A cleaning mode switching method and system based on a multifunctional cleaning robot, and a storage medium

Abstract

The present application relates to a kind of based on multifunctional cleaning robot cleaning mode switching method, system and storage medium, the method includes Q1.multifunctional cleaning robot travels on road, based on airborne camera real-time acquisition road image data information, and the road image data information is carried out Gaussian noise enhancement processing, enhanced road image data information is output;Q2.The enhanced road image data information is classified as training image set and test image set, and the training image set is input first neural network model and is trained and learned, obtains trained first neural network model, and the test training image set is input trained first neural network type and is optimized to image pixel probability distribution.The present application not only can improve the cleaning efficiency of cleaning robot, and according to different garbage kind carries out different cleaning mode switching, guarantees the improvement of cleaning quality.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cleaning robots, in particular to a cleaning mode switching method and system based on a multifunctional cleaning robot and a storage medium. BACKGROUND

[0002] With the continuous progress of social science and technology, the labor cost is increasing, and the application of environmental protection machinery is becoming more and more widely. The main application scene of environmental protection machinery is in public places such as city squares, shopping malls, hospitals, airports, stations, factories, workshops, etc. The requirements of different application scenes for environmental protection machinery are also getting higher and higher. Most of the existing mechanical cleaning equipment has single function, or is only a single sweeping machine, which can only clean large-particle garbage such as fruit peel, paper scraps, leaves, cigarette butts and branches, and has no good cleaning effect on the ground with more small-particle garbage such as dust and dust. Or it is only a single washing machine, which uses a washing brush disc and a water spraying device to wash the ground, and then uses a negative pressure suction disc to collect the sewage, which cannot clean large-particle garbage. There is also a hand-push or manually-driven sweeping and mopping machine that integrates sweeping, dust collection and mopping. These devices are non-modular in design, and each function can only work in a fixed way, and cannot select the required function module for operation according to the actual road conditions, so the flexibility and intelligence are not high. If the area with multiple types of garbage needs to be cleaned, different types of cleaning equipment need to be used to complete the work, which can easily reduce the cleaning efficiency, increase the labor cost, and cannot guarantee the cleaning quality. SUMMARY

[0003] In view of the above problems, the present application provides a cleaning mode switching method and system of a multifunctional cleaning robot and a storage medium, which can not only improve the cleaning efficiency of the cleaning robot, but also switch different cleaning modes according to different types of garbage to improve the cleaning quality.

[0004] In order to achieve the above-mentioned purposes and other related purposes, the technical solutions provided by the present application are as follows:

[0005] A cleaning mode switching method based on a multifunctional cleaning robot, the method comprising:

[0006] Q1. The multifunctional cleaning robot drives on the road, obtains the image data information of the road in real time based on the on-board camera, and performs Gaussian noise enhancement processing on the image data information of the road, and outputs the enhanced road image data information;

[0007] Q2. The enhanced road image data information is classified into a training image set and a test image set, the training image set is input into a first neural network model for training and learning, a trained first neural network model is obtained, and the test training image set is input into the trained first neural network model for optimization of image pixel probability distribution, and an optimized road image pixel matrix data information is output;

[0008] Q3. The optimized road image pixel matrix data information is input into a second neural network model for training and learning, a trained second neural network model is obtained, and the optimized road image pixel matrix data information is input into the trained second neural network model to identify and classify the road garbage, and the garbage quantity is counted according to the classification, and a comprehensive data information of the road garbage is output;

[0009] Q4. Based on the comprehensive data information of the road garbage, a cleaning mode switching decision function G is established to control the cleaning mode switching of the multifunctional cleaning robot.

[0010] Further, in step Q4, the cleaning mode switching decision function G is,

[0011] ,

[0012] wherein x is the comprehensive data information of the road garbage, y is the cleaning efficiency under the current cleaning mode, f i (x,y) is the i-th cleaning planning function, λ i is the corresponding weight value of the i-th cleaning planning function f i (x,y), and M(x) is a road garbage capacity normalization function.

[0013] Further, the road garbage capacity normalization function M(x) is,

[0014] ,

[0015] wherein x is the comprehensive data information of the road garbage, y' is the cleaning efficiency under different cleaning modes, f i (x,y) is the i-th cleaning planning function, λ i is the corresponding weight value of the i-th cleaning planning function f i (x,y).

[0016] Further, in step Q1, the Gaussian noise enhancement processing of the road image data information includes:

[0017] Q11. Based on the road image data information, a grayscale processing is performed to obtain road image matrix data information;

[0018] Q12. Construct a Gaussian distribution noise matrix corresponding to the image matrix data information of the road, add the Gaussian distribution noise matrix to the image matrix data information of the road, and obtain enhanced road image data information.

[0019] Further, in step Q2, the first neural network model includes a mean neuron and a standard deviation neuron, and the inputting of the test training image set into the trained first neural network model for optimization of image pixel probability distribution includes:

[0020] Q21. Input the enhanced road image data information into the mean neuron to obtain mean data information of the image pixels of the road, and input the enhanced road image data information into the standard deviation neuron to obtain standard deviation data information of the image pixels of the road;

[0021] Q22. Based on the mean data information of the image pixels of the road and the standard deviation data information of the image pixels of the road, a Gaussian random number generation algorithm is used to calculate the probability of the image pixels of the road to obtain pixel probability distribution data information of the road image;

[0022] Q23. Based on the pixel probability distribution data information of the road image, a preset image pixel probability threshold is set, if the pixel probability distribution data information of the road image is less than the preset image pixel probability threshold, it is rejected, if the pixel probability distribution data information of the road image is greater than the preset image pixel probability threshold, it is retained, and optimized road image pixel matrix data information is obtained.

[0023] Further, in step Q3, the second neural network model includes a pooling neuron and an interpolation neuron, and the inputting of the optimized road image pixel matrix data information into the trained second neural network model for identification and classification of road garbage includes:

[0024] Q31. Input the optimized road image pixel matrix data information into the pooling neuron to establish a pooling decision function H,

[0025] ,

[0026] wherein, x nn is the image pixel in the nth row and the nth column of the optimized road image pixel matrix, m nn is the decision factor of the image pixel in the nth row and the nth column, and the decision road image pixel matrix data information is obtained.

[0027] Q32. The pixel matrix data information of the selected road image is input into the interpolation neuron, and the pixels of the road image are identified and classified to obtain garbage classification data information.

[0028] To achieve the above object and other related objects, the present application further provides a system for implementing the cleaning mode switching method based on the multifunctional cleaning robot, the system comprising:

[0029] a data acquisition module for acquiring image data information of a road;

[0030] an image enhancement module connected with the data acquisition module for performing enhancement processing on the image data information of the road;

[0031] a first neural network model module connected with the image enhancement module for performing optimization processing on the image data information of the road;

[0032] a second neural network model module connected with the first neural network model module for identifying and classifying the optimized road image data;

[0033] a data decision module connected with the second neural network model for decision of cleaning mode switching.

[0034] Further, the system further comprises a data execution module connected with the data decision module for receiving decision data information of cleaning mode switching and performing switching of the cleaning mode.

[0035] Further, the system further comprises a pre-warning module connected with the data decision module for reminding vehicles and pedestrians to pay attention to avoidance.

[0036] To achieve the above object and other related objects, the present application further provides a computer readable storage medium having stored thereon a computer program programmed or configured to perform any one of the cleaning mode switching methods based on the multifunctional cleaning robot.

[0037] The present application has the following positive effects:

[0038] 1. The present application improves the complexity of the road image by performing Gaussian noise enhancement processing on the image data information of the road, which is closer to the real situation and facilitates the road cleaning operation of the cleaning robot, so as to ensure that the cleaning robot can adapt to different cleaning environments.

[0039] 2. The application optimizes the image data information of the road through a first neural network model, and identifies and classifies the optimized road image data information through a second neural network model, so that the identification of garbage in the road is more accurate, thereby improving the cleaning efficiency of the cleaning robot. At the same time, the cleaning mode of the cleaning robot is switched by cooperating with the decision function, further improving the cleaning quality of the cleaning robot. BRIEF DESCRIPTION OF DRAWINGS

[0040] Figure 1 The figure is a schematic diagram of the method of the application. DETAILED DESCRIPTION

[0041] The exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to help understanding, and should be considered only as exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Also, for the sake of clarity and conciseness, the description below omits the description of well-known functions and structures.

[0042] Embodiment 1: As shown in the figure, a cleaning mode switching method based on a multifunctional cleaning robot, the method comprising: Figure 1

[0043] Q1. The multifunctional cleaning robot drives on the road, obtains the image data information of the road in real time based on the on-board camera, and performs Gaussian noise enhancement processing on the image data information of the road, and outputs the enhanced road image data information;

[0044] Q2. The enhanced road image data information is classified into a training image set and a test image set, and the training image set is input into a first neural network model for training and learning, to obtain a trained first neural network model, and the test training image set is input into the trained first neural network model for optimization of image pixel probability distribution, to output the optimized road image pixel matrix data information;

[0045] Q3. The optimized road image pixel matrix data information is input into a second neural network model for training and learning, to obtain a trained second neural network model, and the optimized road image pixel matrix data information is input into the trained second neural network model to identify and classify the garbage on the road, and according to the classification, the number of garbage is counted, to output the comprehensive data information of the road garbage;

[0046] Q4. Based on the comprehensive data information of the road garbage, a cleaning mode switching decision function G is established to control the cleaning mode switching of the multifunctional cleaning robot.

[0047] ​In the embodiment, in step Q4, the cleaning mode switching decision function G is,

[0048] ,

[0049] wherein x is the comprehensive data information of road garbage, y is the cleaning efficiency under the current cleaning mode, f i (x,y) is the i-th cleaning planning function, λ i is the corresponding weight value of the i-th cleaning planning function f i (x,y), and M(x) is the road garbage capacity normalization function.

[0050] In the embodiment, the road garbage capacity normalization function M(x) is,

[0051] ,

[0052] wherein x is the comprehensive data information of road garbage, y´ is the cleaning efficiency under different cleaning modes, f i (x,y) is the i-th cleaning planning function, λ i is the corresponding weight value of the i-th cleaning planning function f i (x,y).

[0053] In the embodiment, in step Q1, the Gaussian noise enhancement processing of the image data information of the road includes:

[0054] Q11. Based on the image data information of the road, grayscale processing is performed to obtain image matrix data information of the road;

[0055] Q12. Based on the image matrix data information of the road, a noise matrix with Gaussian distribution corresponding thereto is constructed, and the noise matrix with Gaussian distribution is added to the image matrix data information of the road to obtain enhanced road image data information.

[0056] In the embodiment, in step Q2, the first neural network model includes a mean neuron and a standard deviation neuron, and the input of the test training image set into the trained first neural network type for optimization of image pixel probability distribution includes:

[0057] Q21. The enhanced road image data information is input into the mean neuron to obtain mean data information of the image pixels of the road, and the enhanced road image data information is input into the standard deviation neuron to obtain standard deviation data information of the image pixels of the road;

[0058] Q22. Based on the mean data information of the image pixels of the road and the standard deviation data information of the image pixels of the road, a Gaussian random number generation algorithm is used to calculate the probability of the image pixels of the road, and pixel probability distribution data information of the road image is obtained;

[0059] Q23. Based on the pixel probability distribution data information of the road image, a preset image pixel probability threshold is set, if the pixel probability distribution data information of the road image is less than the preset image pixel probability threshold, it is rejected, if the pixel probability distribution data information of the road image is greater than the preset image pixel probability threshold, it is retained, and optimized road image pixel matrix data information is obtained.

[0060] Embodiment 2: Based on the cleaning mode switching method of the multifunctional cleaning robot in embodiment 1, the application is further described and explained as follows.

[0061] As shown in Figure 1 A cleaning mode switching method based on a multifunctional cleaning robot, the method comprising:

[0062] Q1. The multifunctional cleaning robot drives on the road, based on the on-board camera, real-time acquisition of the image data information of the road, and the image data information of the road is processed by Gaussian noise enhancement, and the enhanced road image data information is output;

[0063] Q2. The enhanced road image data information is classified into a training image set and a test image set, and the training image set is input into a first neural network model for training and learning, and a trained first neural network model is obtained, and the test training image set is input into the trained first neural network type for image pixel probability distribution optimization, and the optimized road image pixel matrix data information is output;

[0064] Q3. The optimized road image pixel matrix data information is input into a second neural network model for training and learning, and a trained second neural network model is obtained, and the optimized road image pixel matrix data information is input into the trained second neural network model to identify and classify the road garbage, and the garbage quantity is counted according to the classification, and the comprehensive data information of the road garbage is output;

[0065] Q4. Based on the comprehensive data information of the road garbage, a cleaning mode switching decision function G is established to control the cleaning mode switching of the multifunctional cleaning robot.

[0066] In the embodiment, in step Q3, the second neural network model comprises a pooling neuron and an interpolation neuron, and the input of the optimized road image pixel matrix data information into the trained second neural network model for identifying and classifying the road garbage comprises:

[0067] Q31. The optimized road image pixel matrix data information is input into the pooling neuron to establish a pooling decision function H,

[0068] ,

[0069] wherein x nn is an image pixel in the nth row and the nth column of the optimized road image pixel matrix, m nn is a decision factor of the image pixel in the nth row and the nth column, and the decision road image pixel matrix data information is obtained.

[0070] Q32. The decision road image pixel matrix data information is input into the interpolation neuron to identify and classify the road image pixel, and the garbage classification data information is obtained.

[0071] The application provides a system for implementing the cleaning mode switching method based on the multifunctional cleaning robot, and the system comprises:

[0072] A data acquisition module is configured to acquire image data information of a road.

[0073] An image enhancement module is connected with the data acquisition module and configured to perform enhancement processing on the image data information of the road.

[0074] A first neural network model module is connected with the image enhancement module and configured to perform optimization processing on the image data information of the road.

[0075] A second neural network model module is connected with the first neural network model module and configured to identify and classify the optimized road image data.

[0076] A data decision module is connected with the second neural network model and configured to make a decision on the cleaning mode switching.

[0077] In the embodiment, the system further comprises a data execution module connected with the data decision module, configured to receive the decision data information of the cleaning mode switching and perform the switching of the cleaning mode.

[0078] In the embodiment, the system further comprises a pre-warning module connected with the data decision module, configured to remind vehicles and pedestrians to pay attention to avoidance.

[0079] The application provides a computer readable storage medium, which stores a computer program programmed or configured to perform any one of the cleaning mode switching methods based on the multifunctional cleaning robot.

[0080] Any reference to memory, storage, database, or other medium used in the embodiments provided by the present application can include non-volatile and / or volatile memory. The non-volatile memory can include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. The volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, the RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), Synchlink DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct RAMbus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0081] In summary, the application can improve the cleaning efficiency of the cleaning robot, and switch the cleaning mode according to different garbage types, thereby improving the cleaning quality.

[0082] The above detailed description does not constitute a limitation on the protection scope of the present disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modification, equivalent replacement, and improvement within the spirit and principle of the present disclosure should be included in the protection scope of the present disclosure.

Claims

1. A cleaning mode switching method based on a multifunctional cleaning robot, characterized in that: The method comprises: Q1. A multifunctional cleaning robot travels on a road, acquires real-time road image data using an onboard camera, performs Gaussian noise enhancement on the road image data, and outputs the enhanced road image data; Q2. Classifying the enhanced road image data into a training image set and a test image set, inputting the training image set into a first neural network model for training and learning to obtain a trained first neural network model, and inputting the test image set into the trained first neural network model for optimizing the image pixel probability distribution, thereby outputting optimized road image pixel matrix data. Q3. Input the optimized road image pixel matrix data into a second neural network model for training and learning, thereby obtaining a trained second neural network model. The optimized road image pixel matrix data is then input into the trained second neural network model to identify and classify road trash, calculate the amount of trash based on the classification, and output comprehensive road trash data. Q4. Based on the comprehensive data information of the road garbage, establish a cleaning mode switching decision function G to control the cleaning mode switching of the multifunctional cleaning robot; In step Q4, the cleaning mode switching decision function G is, ， Among them, x is the comprehensive data information of road garbage, y is the cleaning efficiency under the current cleaning mode, and f i (x, y) is the i-th sweep planning function, λ i For the i-th sweep planning function f i The corresponding weight value of (x, y), M(x) is the normalized function of road garbage capacity; In step Q2, the first neural network model includes mean neurons and standard deviation neurons, and the inputting of the test image set into the trained first neural network model to optimize the image pixel probability distribution includes: Q21. Input the enhanced road image data information into the mean neuron to obtain mean data information of the road image pixels, and input the enhanced road image data information into the standard deviation neuron to obtain standard deviation data information of the road image pixels; Q22. Based on the mean data information and the standard deviation data information of the road image pixels, a Gaussian random number generation algorithm is used to calculate the probability of the road image pixels to obtain pixel probability distribution data information of the road image; Q23. Based on the pixel probability distribution data information of the road image, set a preset image pixel probability threshold, and discard the pixel probability distribution data of the road image if it is less than the preset image pixel probability threshold; retain the pixel probability distribution data of the road image if it is greater than the preset image pixel probability threshold, thereby obtaining optimized road image pixel matrix data information; The road garbage capacity normalization function M(x), ， Among them, x is the comprehensive data information of road garbage, y' is the cleaning efficiency under different cleaning modes, and f i (x, y) is the i-th sweep planning function, λ i For the i-th sweep planning function f i The corresponding weight value of (x,y).

2. The cleaning mode switching method based on the multifunctional cleaning robot according to claim 1, characterized in that: In step Q1, performing Gaussian noise enhancement processing on the image data information of the road includes: Q11. Based on the image data information of the road, grayscale processing is performed to obtain the image matrix data information of the road; Q12. Based on the image matrix data information of the road, construct a corresponding Gaussian distribution noise matrix, add the Gaussian distribution noise matrix to the image matrix data information of the road, and obtain enhanced road image data information.

3. The cleaning mode switching method based on the multifunctional cleaning robot according to claim 1, characterized in that: In step Q3, the second neural network model includes pooling neurons and interpolation neurons, and inputting the optimized road image pixel matrix data information into the trained second neural network model to identify and classify road garbage includes: Q31. Input the optimized road image pixel matrix data information into the pooling neuron to establish a pooling decision function H. ， Among them, x nn is the image pixel in the nth row and nth column of the optimized road image pixel matrix, m nn is the decision factor of the image pixel in the nth row and nth column, and the road image pixel matrix data information after the decision is obtained; Q32. Input the selected road image pixel matrix data information into the interpolation neuron, identify and classify the pixels of the road image, and obtain garbage classification data information.

4. A system for implementing the cleaning mode switching method based on a multifunctional cleaning robot according to any one of claims 1 to 3, characterized in that: The system comprises: A data acquisition module is used to acquire image data information of the road; An image enhancement module, connected to the data acquisition module, for enhancing the image data information of the road; A first neural network model module is connected to the image enhancement module and is used to optimize the image data information of the road; a second neural network model module, connected to the first neural network model module, for identifying and classifying the optimized road image data; A data decision module is connected to the second neural network model and is used to make decisions on cleaning mode switching.

5. The system according to claim 4, characterized in that: The system further comprises a data execution module connected to the data decision module, configured to receive decision data information for switching the cleaning mode and perform switching of the cleaning mode.

6. The system according to claim 5, characterized in that: The system also includes an early warning module connected to the data decision module, which is used to remind vehicles and pedestrians to pay attention to avoidance.

7. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program that is programmed or configured to execute the cleaning mode switching method based on the multifunctional cleaning robot as described in any one of claims 1 to 3.

Images