Neural network-based swimming pool cleaning method and device, robot and storage medium

Abstract

The invention is suitable for the technical field of robot cleaning, and provides a swimming pool cleaning method based on a neural network, and the method comprises the steps: recognizing the swimming pool shape of a target swimming pool based on a swimming pool shape recognition model which is trained in advance and is constructed through a neural network when a request for cleaning the target swimming pool is received, and on the basis of the identified swimming pool shape, generating a target cleaning strategy for the underwater robot to clean the target swimming pool, and controlling the underwater robot to clean the target swimming pool according to the target cleaning strategy, so that the cleaning coverage rate, the cleaning efficiency and the cleaning effect of the underwater robot for cleaning the swimming pool are improved, and the user experience is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of robot cleaning, and in particular relates to a swimming pool cleaning method, device, robot and storage medium based on neural network. Background Art

[0002] With the rapid advancement of computer and robotics technology, robots have penetrated into many areas of people's lives, including household cleaning, industrial applications, and special environment operations. In the field of pool cleaning, the application of underwater robots has significantly improved the efficiency and convenience of underwater cleaning.

[0003] At present, although the technology of pool cleaning robots has made significant progress in automation and convenience, its positioning accuracy and communication stability are still the key factors restricting its performance improvement when facing complex underwater environments. The particularity of the underwater environment, such as insufficient light, fast signal attenuation, and water flow interference, greatly increases the difficulty of the robot to achieve accurate navigation and efficient communication. Therefore, the existing pool cleaning robots often use inertial navigation data for navigation, or use random paths for cleaning. Although the traditional inertial navigation data navigation method can guide the robot to move along the preset path to a certain extent, this method depends on the laying and maintenance of external equipment, which not only increases the installation cost, but also limits the flexibility and scope of application of the robot, especially in large or irregularly shaped swimming pools, where the inertial navigation layout becomes complicated and inefficient. On the other hand, although the robot using random path cleaning does not rely on external navigation equipment, its cleaning process lacks systematic planning, which easily leads to cleaning blind spots, especially near the corners, edges or obstacles of the swimming pool, and the cleaning coverage is difficult to guarantee, which not only affects the cleaning effect of the swimming pool, but also may extend the cleaning cycle and reduce the user experience. Therefore, a new swimming pool cleaning method is urgently needed to improve the coverage and efficiency of swimming pool cleaning. Summary of the invention

[0004] The object of the present invention is to provide a swimming pool cleaning method, device, robot and storage medium based on a neural network, aiming to solve the problems of low swimming pool cleaning coverage and poor effect caused by the prior art.

[0005] In one aspect, the present invention provides a swimming pool cleaning method based on a neural network, the method comprising the following steps:

[0006] When a request for cleaning a target swimming pool is received, the swimming pool shape of the target swimming pool is identified based on a pre-trained swimming pool shape recognition model constructed using a neural network;

[0007] Based on the identified swimming pool shape, generating a target cleaning strategy for the underwater robot to clean the target swimming pool;

[0008] The underwater robot is controlled to clean the target swimming pool according to the target cleaning strategy.

[0009] Preferably, the swimming pool shape recognition model is a model constructed using a convolutional neural network.

[0010] Preferably, before identifying the swimming pool shape of the target swimming pool based on the pre-trained swimming pool shape recognition model, the method further includes:

[0011] Obtaining a training data set, wherein each sample in the training data set is a path graph;

[0012] Preprocessing the training data set to obtain the true shape of the swimming pool corresponding to each sample in the training data set;

[0013] The swimming pool shape recognition model is trained using the preprocessed training data set, and the model parameters of the swimming pool shape recognition model are adjusted through a back propagation algorithm until the error between the predicted shape of the swimming pool corresponding to each sample in the training data set predicted by the swimming pool shape recognition model and the corresponding true shape is minimized.

[0014] Preferably, the step of identifying the swimming pool shape of the target swimming pool based on a pre-trained swimming pool shape recognition model comprises:

[0015] collecting cleaning path data of the underwater robot in real time during the cleaning of the target swimming pool;

[0016] The swimming pool shape is identified according to the cleaning path data using the swimming pool shape recognition model.

[0017] Preferably, based on the identified swimming pool shape, the step of generating a target cleaning strategy for the underwater robot to clean the target swimming pool comprises:

[0018] Determining a cleaning path for the underwater robot to clean the target swimming pool according to the shape of the swimming pool, wherein the cleaning path is one path or a combination of multiple different paths;

[0019] Controlling the underwater robot to monitor the water quality parameters of the target swimming pool in real time during the cleaning process;

[0020] The cleaning parameters of the underwater robot during the cleaning process are adjusted in real time according to the water quality parameters.

[0021] In another aspect, the present invention provides a swimming pool cleaning device based on a neural network, the device comprising:

[0022] A shape recognition unit, configured to recognize the pool shape of the target pool based on a pre-trained pool shape recognition model constructed using a neural network when a request for cleaning the target pool is received;

[0023] A strategy generating unit, configured to generate a target cleaning strategy for the underwater robot to clean the target swimming pool based on the identified swimming pool shape;

[0024] The swimming pool cleaning unit is used to control the underwater robot to clean the target swimming pool according to the target cleaning strategy.

[0025] Preferably, the device further comprises:

[0026] A data acquisition unit, used to acquire a training data set, wherein each sample in the training data set is a path graph;

[0027] A data processing unit, used for preprocessing the training data set to obtain the real shape of the swimming pool corresponding to each sample in the training data set;

[0028] A model training unit is used to train the swimming pool shape recognition model using the preprocessed training data set, and adjust the model parameters of the swimming pool shape recognition model through a back propagation algorithm until the error between the predicted shape of the swimming pool corresponding to each sample in the training data set predicted by the swimming pool shape recognition model and the corresponding true shape is minimized.

[0029] Preferably, the strategy generation unit includes:

[0030] A cleaning path determination unit, used to determine a cleaning path for the underwater robot to clean the target swimming pool according to the shape of the swimming pool, wherein the cleaning path is one path or a combination of multiple different paths;

[0031] A water quality parameter monitoring unit, used to control the underwater robot to monitor the water quality parameters of the target swimming pool in real time during the cleaning process;

[0032] A cleaning parameter adjustment unit is used to adjust the cleaning parameters of the underwater robot during the cleaning process in real time according to the water quality parameters.

[0033] On the other hand, the present invention also provides an underwater robot, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps described in the above-mentioned neural network-based swimming pool cleaning method are implemented.

[0034] On the other hand, the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps described in the above-mentioned neural network-based swimming pool cleaning method are implemented.

[0035] When the present invention receives a request to clean a target swimming pool, the pool shape of the target swimming pool is identified based on a pre-trained swimming pool shape recognition model constructed using a neural network, and based on the identified swimming pool shape, a target cleaning strategy for an underwater robot to clean the target swimming pool is generated, and the underwater robot is controlled to clean the target swimming pool according to the target cleaning strategy, thereby improving the cleaning coverage, cleaning efficiency and cleaning effect of the underwater robot for cleaning the swimming pool, thereby improving the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 is a flow chart of an implementation of a swimming pool cleaning method based on a neural network provided in Embodiment 1 of the present invention;

[0037] Figure 2 is a schematic diagram of a simulated cleaning path provided in Embodiment 1 of the present invention;

[0038] Figure 3 is a schematic diagram of a bow-shaped / spiral-shaped / special-shaped cleaning path provided in the first embodiment of the present invention;

[0039] Figure 4 is a schematic diagram of the structure of a swimming pool cleaning device based on a neural network provided in the second embodiment of the present invention;

[0040] Figure 5 is a schematic diagram of the structure of a swimming pool cleaning device based on a neural network provided in Embodiment 3 of the present invention;

[0041] Figure 6 It is a schematic diagram of the structure of the underwater robot provided in the fourth embodiment of the present invention. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0043] The specific implementation of the present invention is described in detail below in conjunction with specific embodiments:

[0044] Embodiment 1:

[0045] Figure 1The implementation process of the swimming pool cleaning method based on neural network provided in the first embodiment of the present invention is shown. For the convenience of description, only the part related to the embodiment of the present invention is shown, which is described in detail as follows:

[0046] In step S101, when a request to clean a target swimming pool is received, the swimming pool shape of the target swimming pool is identified based on a pre-trained swimming pool shape recognition model constructed using a neural network.

[0047] The embodiment of the present invention is applicable to an underwater robot for swimming pool cleaning operations, which supports three modes: pool bottom cleaning, wall climbing cleaning, and waterline cleaning, so as to help users automatically clean the dirt on the bottom and wall of the swimming pool and reduce the workload of human intervention. In the embodiment of the present invention, when a request for cleaning a target swimming pool is received, the swimming pool shape of the target swimming pool is identified based on a pre-trained swimming pool shape recognition model, wherein the swimming pool shape recognition model is a model constructed using a neural network, and the swimming pool shape includes regular shapes and irregular shapes, regular shapes include but are not limited to rectangles, circles, and ellipses, and irregular shapes include but are not limited to triangles, L shapes, and wave shapes that simulate the wave shape of the ocean. In addition, the request for cleaning the target swimming pool can be triggered by the user through an electronic terminal (such as a mobile phone, a personal computer) wirelessly connected to the underwater robot, or by starting a cleaning button set in the underwater robot. The triggering conditions for the cleaning request of the target swimming pool are not specifically limited here.

[0048] In a feasible embodiment, a convolutional neural network (CNN) is used to construct a pool shape recognition model for an underwater robot to identify the pool shape of a target pool, thereby improving the intelligence of the underwater robot and the subsequent cleaning effect.

[0049] In another feasible embodiment, before recognizing the pool shape of the target pool based on the pre-trained pool shape recognition model, the pool shape recognition model is trained by the following steps:

[0050] ① Obtain a training data set, where each sample in the training data set is a path graph;

[0051] In the embodiment of the present invention, a large number of cleaning paths of underwater robots for a complete cleaning of swimming pools of different shapes and sizes are collected, and a training data set is constructed based on the path graphs corresponding to these cleaning paths. The cleaning path can be a historical cleaning path when the user uses the underwater robot to clean the swimming pool. If the amount of historical cleaning path data is insufficient, MATLAB or other simulation software is used to simulate the cleaning process of the underwater robot in swimming pools of different shapes to generate a simulated cleaning path containing the robot's movement trajectory (such as Figure 2 As shown in the figure), the training data set is then composed of the path graphs corresponding to the historical cleaning paths and the simulated cleaning paths to enhance the robustness and practicality of the model.

[0052] ②Preprocess the training data set to obtain the true shape of the swimming pool corresponding to each sample in the training data set;

[0053] In an embodiment of the present invention, each sample in the training data set is preprocessed, including data standardization, normalization, labeling and other steps to adapt to the input requirements of the swimming pool shape recognition model constructed using CNN, and a label is assigned to each sample in the training data set, which represents the true shape of the swimming pool corresponding to the sample, wherein the label can be encoded with numbers (for example, label "1" represents a matrix, label "2" represents a circle, etc.), or with letters, and the encoding form of the label is not specifically limited herein.

[0054] ③ Use the preprocessed training data set to train the swimming pool shape recognition model, and adjust the model parameters of the swimming pool shape recognition model through the back propagation algorithm until the error between the predicted shape of the swimming pool corresponding to each sample in the training data set predicted by the swimming pool shape recognition model and the corresponding true shape is minimized.

[0055] In an embodiment of the present invention, the preprocessed training data set is input into the swimming pool shape recognition model to perform model training. Specifically, during the forward propagation process, the swimming pool shape recognition model will traverse each sample in the training data set, predict the shape of the swimming pool corresponding to each sample, and output the corresponding predicted shape (such as rectangle, circle, irregular shape, etc.). Subsequently, according to the error between each predicted shape and the corresponding true shape (usually calculated by the loss function, such as cross entropy loss, mean square error, etc.), the back propagation algorithm is used to adjust the parameters of the swimming pool shape recognition model. Here, the back propagation algorithm gradually adjusts the weights and biases within the model through gradient descent or other optimization algorithms to minimize the error (i.e., loss) between the predicted shape and the true shape. This process will be repeated for multiple iterations until the preset number of iterations is reached, the loss value is reduced to a sufficiently small value, or the model performance is no longer significantly improved on the validation set. Through such a training process, the swimming pool shape recognition model can learn the feature representations of swimming pools of different shapes, thereby improving the accuracy of swimming pool shape recognition.

[0056] Through the above steps ① to ③, the swimming pool shape recognition model is trained, thereby improving the accuracy of model recognition.

[0057] In another feasible embodiment, when identifying the pool shape of the target swimming pool based on a pre-trained swimming pool shape recognition model, the cleaning path data of the underwater robot during the cleaning of the target swimming pool is collected in real time, and the pool shape is identified based on the cleaning path data through the swimming pool shape recognition model.

[0058] In an embodiment of the present invention, when the underwater robot performs the task of cleaning the target swimming pool, it will collect its underwater cleaning path data, that is, the motion trajectory, in real time, which includes but is not limited to the position coordinates and timestamp of the underwater robot in the target swimming pool. After that, the cleaning path data is preprocessed to remove noise, fill missing values, smooth the path, etc., to ensure the accuracy and reliability of the data. Then, the preprocessed cleaning path data is converted into a format suitable for input to the swimming pool shape recognition model. Specifically, according to the corresponding path points in the cleaning path data, the cleaning path data is converted into a path image or other forms of spatial representation. Finally, the converted path image or other forms of spatial representation is input into the swimming pool shape recognition model. The swimming pool shape recognition model first extracts features that can reflect the shape of the swimming pool (for example, boundary shape, length, curvature change, inflection point distribution, etc.), and then outputs the recognition result of the swimming pool shape, such as rectangle, circle, irregular shape, etc., through the forward propagation process based on these features.

[0059] In step S102, based on the identified swimming pool shape, a target cleaning strategy for the underwater robot to clean the target swimming pool is generated.

[0060] In an embodiment of the present invention, based on the identified shape of the swimming pool, a target cleaning strategy is generated for the underwater robot to clean the target swimming pool. The target cleaning strategy includes a series of planning, execution and adjustment measures taken to achieve specific cleaning goals (such as comprehensive, efficient and energy-saving cleaning of the swimming pool).

[0061] In a feasible embodiment, the generation of the target cleaning strategy is achieved through the following steps:

[0062] (1) determining a cleaning path for the underwater robot to clean the target swimming pool according to the shape of the swimming pool, wherein the cleaning path is one path or a combination of multiple different paths;

[0063] In an embodiment of the present invention, based on the identified shape of the swimming pool, the underwater robot can automatically plan the optimal cleaning path to ensure full coverage of the bottom and walls of the swimming pool while avoiding repeated cleaning and missed areas, wherein the cleaning path is one path or a combination of multiple different paths.

[0064] In a feasible embodiment, for a rectangular swimming pool, a straight scanning path parallel to the long side or short side of the swimming pool is designed in combination with the length and width data of the target swimming pool measured by the underwater robot. Figure 3 (a) to ensure full coverage of the pool bottom and side walls, or an alternating vertical and horizontal bow-shaped path (i.e., on the basis of the bow-shaped path, the vertical and horizontal modes of the bow-shaped path are switched each time, such as the nth time is a vertical bow-shaped path, and the n+1th time is a horizontal bow-shaped path) to reduce the missed gaps between parallel paths; for circular or elliptical pools, combined with the diameter or major / minor axis of the target pool measured by the underwater robot, a circular or spiral path is used to expand from the center outward (such as Figure 3 (b) to reduce repeated cleaning and omissions, or use a spiral path to clean after the bow-shaped path to remove dead corners that the bow-shaped path fails to cover; for irregularly shaped swimming pools (such as Figure 3 For the special-shaped swimming pool shown in (c), you can first use a bow-shaped path (red) to clean the edges and corners along the edge of the pool (green), or use the robot's autonomous navigation and obstacle avoidance functions to dynamically plan the cleaning path to ensure that every corner is cleaned.

[0065] (2) Control the underwater robot to monitor the water quality parameters of the target swimming pool in real time during the cleaning process;

[0066] In the embodiment of the present invention, this is the execution and monitoring part of the target cleaning strategy. By controlling the underwater robot to monitor the water quality parameters of the target swimming pool in real time during the cleaning process, the current water quality status of the target swimming pool can be timely understood, thereby providing a basis for subsequent cleaning parameter adjustments, wherein the water quality parameters include but are not limited to the chlorine content, pH value, suspended matter, algae or other impurities in the water. Density.

[0067] (3) Adjust the cleaning parameters of the underwater robot in real time during the cleaning process according to water quality parameters.

[0068] In an embodiment of the present invention, based on the water quality parameters monitored in real time, the underwater robot can automatically adjust its cleaning parameters (such as suction, scrubbing intensity, cleaning speed, etc.) during the cleaning process. This is the dynamic adjustment part of the target cleaning strategy. By continuously adjusting the cleaning parameters, the underwater robot can respond quickly to different water quality conditions, thereby ensuring the best cleaning effect.

[0069] The target cleaning strategy is generated through the above steps (1) to (3), thereby achieving efficient, comprehensive and energy-saving cleaning of the swimming pool through intelligent planning, execution and adjustment.

[0070] In step S103, the underwater robot is controlled to clean the target swimming pool according to the target cleaning strategy.

[0071] In an embodiment of the present invention, when controlling the underwater robot to clean the target swimming pool according to the target cleaning strategy, specifically, the underwater robot travels in the target swimming pool according to the cleaning path in the strategy, and cleans the swimming pool according to the cleaning parameters corresponding to the current water quality conditions. During the cleaning operation, the underwater robot adjusts parameters such as the rotation speed, suction force, and travel speed of the brush to adapt to the cleaning needs of different areas and different water qualities. For example, the rotation speed and suction force of the brush are increased in areas with more sediments at the bottom of the swimming pool, and the travel speed is reduced in smooth areas such as the wall of the swimming pool to improve the cleaning effect.

[0072] In a feasible embodiment, when the underwater robot detects that the water quality parameter reaches a preset cleaning standard or the cleaning time (number of times) reaches a preset time (number of times) threshold, the cleaning operation is stopped.

[0073] In an embodiment of the present invention, when a request to clean a target swimming pool is received, the shape of the target swimming pool is identified based on a pre-trained swimming pool shape recognition model constructed using a neural network, and based on the identified swimming pool shape, a target cleaning strategy is generated for an underwater robot to clean the target swimming pool, and the underwater robot is controlled to clean the target swimming pool according to the target cleaning strategy, thereby improving the cleaning coverage, cleaning efficiency and cleaning effect of the underwater robot cleaning the swimming pool, thereby improving the user experience.

[0074] Embodiment 2:

[0075] Figure 4 The structure of a swimming pool cleaning device based on a neural network provided in the second embodiment of the present invention is shown. For the sake of convenience, only the parts related to the embodiment of the present invention are shown, including:

[0076] The shape recognition unit 41 is used to recognize the pool shape of the target pool based on a pre-trained pool shape recognition model constructed using a neural network when a request for cleaning the target pool is received;

[0077] A strategy generating unit 42, for generating a target cleaning strategy for the underwater robot to clean the target swimming pool based on the identified swimming pool shape;

[0078] The swimming pool cleaning unit 43 is used to control the underwater robot to clean the target swimming pool according to the target cleaning strategy.

[0079] In the embodiment of the present invention, each unit of the pool cleaning device based on the neural network can be implemented by a corresponding hardware or software unit, and each unit can be an independent software or hardware unit, or can be integrated into a software or hardware unit, which is not intended to limit the present invention. Specifically, the implementation of each unit can refer to the description of the aforementioned embodiment 1, which will not be repeated here.

[0080] Embodiment three:

[0081] Figure 5 The structure of a swimming pool cleaning device based on a neural network provided in the third embodiment of the present invention is shown. For the sake of convenience, only the parts related to the embodiment of the present invention are shown, including:

[0082] A data acquisition unit 51 is used to acquire a training data set, wherein each sample in the training data set is a path graph;

[0083] A data processing unit 52 is used to pre-process the training data set to obtain the true shape of the swimming pool corresponding to each sample in the training data set;

[0084] A model training unit 53 is used to train the swimming pool shape recognition model using the preprocessed training data set, and adjust the model parameters of the swimming pool shape recognition model through a back propagation algorithm until the error between the predicted shape of the swimming pool corresponding to each sample in the training data set predicted by the swimming pool shape recognition model and the corresponding true shape is minimized;

[0085] A shape recognition unit 54 is used to recognize the shape of the target swimming pool based on a pre-trained swimming pool shape recognition model constructed using a neural network when a request for cleaning the target swimming pool is received;

[0086] A strategy generating unit 55, for generating a target cleaning strategy for the underwater robot to clean the target swimming pool based on the identified swimming pool shape;

[0087] The swimming pool cleaning unit 56 is used to control the underwater robot to clean the target swimming pool according to the target cleaning strategy.

[0088] Preferably, the strategy generation unit 55 includes:

[0089] A cleaning path determination unit, used to determine a cleaning path for the underwater robot to clean a target swimming pool according to the shape of the swimming pool, wherein the cleaning path is one path or a combination of multiple different paths;

[0090] A water quality parameter monitoring unit is used to control the underwater robot to monitor the water quality parameters of the target swimming pool in real time during the cleaning process;

[0091] The cleaning parameter adjustment unit is used to adjust the cleaning parameters of the underwater robot during the cleaning process in real time according to the water quality parameters.

[0092] In another embodiment, the shape recognition unit 54 includes:

[0093] A path data acquisition unit is used to collect cleaning path data of the underwater robot in the process of cleaning the target swimming pool in real time;

[0094] The shape recognition subunit is used to recognize the shape of the swimming pool through the swimming pool shape recognition model according to the cleaning path data.

[0095] In another preferred embodiment, the swimming pool shape recognition model is a model constructed using a convolutional neural network.

[0096] In the embodiment of the present invention, each unit of the pool cleaning device based on the neural network can be implemented by a corresponding hardware or software unit, and each unit can be an independent software or hardware unit, or can be integrated into a software or hardware unit, which is not intended to limit the present invention. Specifically, the implementation of each unit can refer to the description of the aforementioned embodiment 1, which will not be repeated here.

[0097] Embodiment 4:

[0098] Figure 6 The structure of an underwater robot provided by a fourth embodiment of the present invention is shown. For ease of description, only the parts related to the embodiment of the present invention are shown.

[0099] The underwater robot 6 of the embodiment of the present invention includes a processor 60, a memory 61, and a computer program 62 stored in the memory 61 and executable on the processor 60. When the processor 60 executes the computer program 62, the steps in the above-mentioned neural network-based swimming pool cleaning method embodiment are implemented, for example Figure 1 Alternatively, when the processor 60 executes the computer program 62, the functions of each unit in the above-mentioned device embodiments are realized, for example Figure 4 Function of the unit shown.

[0100] In an embodiment of the present invention, when a request to clean a target swimming pool is received, the pool shape of the target swimming pool is identified based on a pre-trained swimming pool shape recognition model constructed using a neural network, and based on the identified swimming pool shape, a target cleaning strategy is generated for an underwater robot to clean the target swimming pool, and the underwater robot is controlled to clean the target swimming pool according to the target cleaning strategy, thereby improving the cleaning coverage, cleaning efficiency and cleaning effect of the underwater robot cleaning the swimming pool, thereby improving the user experience.

[0101] The steps implemented when the processor 60 in the underwater robot 6 of the embodiment of the present invention executes the computer program 62 to implement the neural network-based swimming pool cleaning method can refer to the description of the aforementioned method embodiment, which will not be repeated here.

[0102] Embodiment five:

[0103] In an embodiment of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the above-mentioned neural network-based swimming pool cleaning method embodiment are implemented, for example, Figure 1 Alternatively, when the computer program is executed by a processor, the functions of each unit in the above-mentioned device embodiments are realized, for example Figure 4 Function of the unit shown.

[0104] In an embodiment of the present invention, when a request to clean a target swimming pool is received, the shape of the target swimming pool is identified based on a pre-trained swimming pool shape recognition model constructed using a neural network, and based on the identified swimming pool shape, a target cleaning strategy is generated for an underwater robot to clean the target swimming pool, and the underwater robot is controlled to clean the target swimming pool according to the target cleaning strategy, thereby improving the cleaning coverage, cleaning efficiency and cleaning effect of the underwater robot cleaning the swimming pool, thereby improving the user experience.

[0105] The computer-readable storage medium of the embodiment of the present invention may include any entity or device or recording medium capable of carrying computer program code, for example, ROM / RAM, magnetic disk, optical disk, flash memory and other memories.

[0106] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A swimming pool cleaning method based on neural network, characterized in that: The method comprises the following steps: When a request for cleaning a target swimming pool is received, the swimming pool shape of the target swimming pool is identified based on a pre-trained swimming pool shape recognition model constructed using a neural network; Based on the identified swimming pool shape, generating a target cleaning strategy for the underwater robot to clean the target swimming pool; The underwater robot is controlled to clean the target swimming pool according to the target cleaning strategy.

2. The method according to claim 1, characterized in that The swimming pool shape recognition model is a model constructed using a convolutional neural network.

3. The method according to claim 1, characterized in that Before identifying the swimming pool shape of the target swimming pool based on the pre-trained swimming pool shape recognition model, the method further includes: Obtaining a training data set, wherein each sample in the training data set is a path graph; Preprocessing the training data set to obtain the true shape of the swimming pool corresponding to each sample in the training data set; The swimming pool shape recognition model is trained using the preprocessed training data set, and the model parameters of the swimming pool shape recognition model are adjusted through a back propagation algorithm until the error between the predicted shape of the swimming pool corresponding to each sample in the training data set predicted by the swimming pool shape recognition model and the corresponding true shape is minimized.

4. The method according to claim 1, characterized in that The step of identifying the swimming pool shape of the target swimming pool based on the pre-trained swimming pool shape recognition model comprises: collecting cleaning path data of the underwater robot in real time during the cleaning of the target swimming pool; The swimming pool shape is identified according to the cleaning path data using the swimming pool shape recognition model.

5. The method according to claim 1, characterized in that The step of generating a target cleaning strategy for the underwater robot to clean the target swimming pool based on the identified swimming pool shape comprises: Determining a cleaning path for the underwater robot to clean the target swimming pool according to the shape of the swimming pool, wherein the cleaning path is one path or a combination of multiple different paths; Controlling the underwater robot to monitor the water quality parameters of the target swimming pool in real time during the cleaning process; The cleaning parameters of the underwater robot during the cleaning process are adjusted in real time according to the water quality parameters.

6. A swimming pool cleaning device based on neural network, characterized in that: The device comprises: A shape recognition unit, configured to recognize the pool shape of the target pool based on a pre-trained pool shape recognition model constructed using a neural network when a request for cleaning the target pool is received; A strategy generating unit, configured to generate a target cleaning strategy for the underwater robot to clean the target swimming pool based on the identified swimming pool shape; The swimming pool cleaning unit is used to control the underwater robot to clean the target swimming pool according to the target cleaning strategy.

7. The device according to claim 6, characterized in that The device also includes: A data acquisition unit, used to acquire a training data set, wherein each sample in the training data set is a path graph; A data processing unit, used for preprocessing the training data set to obtain the real shape of the swimming pool corresponding to each sample in the training data set; A model training unit is used to train the swimming pool shape recognition model using the preprocessed training data set, and adjust the model parameters of the swimming pool shape recognition model through a back propagation algorithm until the error between the predicted shape of the swimming pool corresponding to each sample in the training data set predicted by the swimming pool shape recognition model and the corresponding true shape is minimized.

8. The device according to claim 6, characterized in that The strategy generation unit comprises: A cleaning path determination unit, used to determine a cleaning path for the underwater robot to clean the target swimming pool according to the shape of the swimming pool, wherein the cleaning path is one path or a combination of multiple different paths; A water quality parameter monitoring unit, used to control the underwater robot to monitor the water quality parameters of the target swimming pool in real time during the cleaning process; A cleaning parameter adjustment unit is used to adjust the cleaning parameters of the underwater robot during the cleaning process in real time according to the water quality parameters.

9. An underwater robot comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 5 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images