Carpet cleaning method and device, cleaning robot and storage medium

Abstract

The embodiment of the invention discloses a carpet cleaning method and device, a cleaning robot and a storage medium. The method is used for reducing the probability that a small carpet is dragged or rolled up by a cleaning robot and comprises the steps that a target image of the ground to be cleaned is acquired through image acquisition equipment; carrying out carpet identification processing on the to-be-cleaned ground according to the target image to obtain a carpet identification result; when the carpet recognition result is that the small carpet is laid on the ground to be cleaned, the cleaning mode of the cleaning robot is adjusted to be a carpet cleaning mode, and at least one small carpet cleaning path is generated according to a preset small carpet path planning algorithm; and performing carpet cleaning work on the small carpet according to the at least one small carpet cleaning path.

Description

Technical Field

[0001] This invention relates to the field of robotics, and more particularly to a carpet cleaning method, apparatus, cleaning robot, and storage medium. Background Technology

[0002] With the development of technology, cleaning robots have gradually evolved from devices that only have the function of sweeping to multi-functional cleaning devices that can sweep, mop, and clean carpets.

[0003] In existing technology, when a cleaning robot comes into contact with a carpet while cleaning a room, it can automatically identify the carpet, adjust the cleaning mode to carpet cleaning mode, and clean the carpet along a predetermined cleaning path (such as a pre-set large room cleaning path).

[0004] However, because existing cleaning robots only switch to carpet cleaning mode when they come into contact with the carpet, the mop or roller brush cannot be lifted in time when the cleaning robot is cleaning a small carpet. When cleaning a small carpet (lightweight carpet), the mop (mop or roller brush) may pass by the edge of the small carpet after the cleaning robot is on the carpet, causing the small carpet to be dragged or rolled up by the cleaning robot. Summary of the Invention

[0005] This invention provides a carpet cleaning method, apparatus, cleaning robot, and storage medium, which can reduce the probability that small carpets are dragged or rolled up by the cleaning robot when cleaning them.

[0006] In a first aspect, embodiments of the present invention provide a carpet cleaning method, wherein the cleaning robot is equipped with an image acquisition device, and the carpet cleaning method includes: The target image of the ground to be cleaned is acquired using the image acquisition device; Based on the target image, carpet recognition processing is performed on the floor to be cleaned to obtain carpet recognition results; When the carpet identification result indicates that a small carpet is laid on the floor to be cleaned, the cleaning robot's cleaning mode is adjusted to carpet cleaning mode, and at least one small carpet cleaning path is generated according to the preset small carpet path planning algorithm. Carpet cleaning is performed on the small carpet according to at least one of the small carpet cleaning paths.

[0007] Secondly, embodiments of the present invention also provide a carpet cleaning device, which is deployed in a cleaning robot. The cleaning robot is equipped with an image acquisition device, and the carpet cleaning device includes: The transceiver unit is used to acquire a target image of the ground to be cleaned through the image acquisition device; The processing unit is used to perform carpet recognition processing on the ground to be cleaned based on the target image to obtain carpet recognition results; when the carpet recognition results indicate that the ground to be cleaned is covered with a small carpet, the cleaning robot's cleaning mode is adjusted to carpet cleaning mode, and at least one small carpet cleaning path is generated according to a preset small carpet path planning algorithm; and carpet cleaning is performed on the small carpet according to at least one of the small carpet cleaning paths.

[0008] Thirdly, embodiments of the present invention also provide a cleaning robot, which includes a memory and a processor. The memory stores a computer program, the cleaning robot is equipped with an image acquisition device, and the processor executes the computer program to implement the above-described method.

[0009] Fourthly, embodiments of the present invention also provide a computer-readable storage medium storing a computer program, the computer program including program instructions that, when executed by a processor, can implement the above-described method.

[0010] Compared with the prior art, the embodiments of the present invention have the following advantages: On the one hand, before cleaning the floor, this embodiment of the invention first determines whether a carpet is laid on the floor by acquiring an image of the floor to be cleaned. If a small carpet is laid, the cleaning mode is adjusted to carpet cleaning mode in advance to ensure that the mop is lifted in advance before the small carpet is placed on it, thus avoiding contact between the mop and the small carpet and reducing the probability of the small carpet being dragged or rolled up by the cleaning robot. On the other hand, before cleaning the small carpet, this embodiment of the invention generates at least one small carpet cleaning path according to a preset small carpet path planning algorithm. There is no need to use a predetermined cleaning path for carpet cleaning. The cleaning robot can adjust its driving position according to the generated small carpet cleaning path to prevent the robot's edge from going beyond the edge of the small carpet, thereby further preventing the small carpet from being dragged or rolled up by the cleaning robot. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 A schematic flowchart of a carpet cleaning method provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of a rectangular carpet path planning provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of a circular carpet path planning provided in an embodiment of the present invention; Figure 4 A schematic diagram of a detour trajectory provided for an embodiment of the present invention; Figure 5 A schematic flowchart of a carpet cleaning method provided in another embodiment of the present invention; Figure 6 A schematic diagram of a cleaning robot performing a carpet-wrapping operation according to an embodiment of the present invention; Figure 7 This is a schematic diagram of a rectangular carpet path planning provided in an embodiment of the present invention; Figure 8 A schematic diagram of a circular carpet path planning provided in an embodiment of the present invention; Figure 9 A schematic block diagram of a carpet cleaning device provided in an embodiment of the present invention; Figure 10 This is a schematic block diagram of a cleaning robot provided in an embodiment of the present invention. Detailed Implementation

[0013] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0014] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0015] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0016] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0017] This invention provides a carpet cleaning method, apparatus, cleaning robot, and storage medium.

[0018] The entity performing this carpet cleaning method can be the carpet cleaning device provided in the embodiments of the present invention, or a cleaning robot that integrates the carpet cleaning device. The carpet cleaning device can be implemented in hardware or software.

[0019] In this invention, the cleaning robot by default includes components found in conventional cleaning robots, such as a shell, wheels, sweeping structure components, water tank, mopping structure components, and control module, and also includes at least an image acquisition device, which can specifically be a camera. The image acquisition device is installed in the front and / or rear areas of the cleaning robot. The front area includes the front, left, and right front areas of the cleaning robot, and the rear area includes the rear, left, and right rear areas of the cleaning robot. The cleaning robot can move forward and backward in any of the following directions: front, left, right, rear, left, and right rear.

[0020] The specific process of this embodiment of the invention is described below, taking a cleaning robot as the execution subject as an example.

[0021] Figure 1 This is a schematic flowchart of the carpet cleaning method provided in an embodiment of the present invention. Figure 1 As shown, the method includes the following steps S110-S140.

[0022] S110. Obtain the target image of the ground to be cleaned through the image acquisition device.

[0023] In this embodiment, during the cleaning process of the cleaning robot in the target cleaning area according to the predetermined path, it is necessary to continuously collect images of the ground to be cleaned through the image acquisition device. The ground to be cleaned is the ground located in the direction of the cleaning robot's movement and at a predetermined distance (such as 30cm) from the cleaning robot. The predetermined cleaning path is the cleaning path set by the cleaning robot for cleaning the target cleaning area.

[0024] Furthermore, in some embodiments, the cleaning robot is also equipped with supplementary lighting devices, such as supplementary lights. Before taking a picture of the ground to be cleaned, the cleaning robot will also detect the brightness of the surrounding ambient light. When the light is sufficient, the supplementary light will be turned off, and when the light is insufficient, the supplementary light will be turned on. The light source provided by the supplementary light will illuminate the ground to be cleaned to ensure that a clear image of the ground to be cleaned is obtained.

[0025] In some embodiments, the cleaning robot is equipped with an ultrasonic sensor on its bottom, which is used to identify the ground material on which the cleaning robot is currently located, such as floor material or carpet material.

[0026] Because the cleaning robot is constantly moving during operation, it may experience carpet recognition anomalies. In such cases, even if the floor to be cleaned has a carpet, the robot might misidentify it as having no carpet. In this situation, the robot will apply the carpet while in floor cleaning mode. During this process, the ultrasonic sensor on the robot's bottom detects the carpet material upon contact with it. When the floor material is carpet, the cleaning robot will retreat a preset distance (e.g., 35cm) to position itself at the image acquisition point of the floor to be cleaned, and generate a carpet detection command. The cleaning robot will then respond to this command by remaining at the image acquisition point for a preset duration (e.g., 2 seconds). While stationary, it will acquire the target image at this image acquisition point using the image acquisition device. This target image will then be used for carpet recognition processing of the floor to be cleaned. This image acquisition point is the optimal acquisition point for the cleaning robot to capture the image of the floor to be cleaned, resulting in the clearest target image.

[0027] S120. Perform carpet recognition processing on the ground to be cleaned based on the target image to obtain carpet recognition results.

[0028] In this embodiment, in order to clean the floor to be cleaned using an appropriate cleaning method, before cleaning the floor to be cleaned, it is necessary to perform carpet recognition processing on the floor to be cleaned using the collected target image and obtain the corresponding carpet recognition result. The carpet recognition result includes whether the floor to be cleaned has a small carpet, a large carpet, or no carpet.

[0029] In some embodiments, carpet identification processing of the floor to be cleaned is performed through the following steps: The target image is preprocessed to obtain a processed image. The preprocessing includes size scaling and pixel normalization. The processed image is then input into a preset carpet recognition model for carpet recognition processing to obtain the carpet recognition result.

[0030] In this embodiment of the invention, the carpet recognition model is a convolutional neural network (CNN) model, specifically a deep convolutional neural network model, such as a residual neural network (ResNet) model, a visual geometry group (VGG) model, or a lightweight (MobileNet) model.

[0031] Specifically, the target image is first scaled to a fixed size, such as 224x224 pixels, to fit the input layer of the CNN model. Pixel normalization is then performed on the target image, normalizing the pixel values ​​of each pixel to the range of (0, 1) to obtain the processed image, thus improving the model's stability. The processed image is then input into the CNN model through its input layer. High-level features of the image are extracted through the convolutional and pooling layers of the CNN model. Finally, the extracted high-level features are input into the fully connected layer of the CNN model to obtain the probability values ​​of various types of carpets. The carpet recognition result is then determined based on the probability values ​​of each type of carpet.

[0032] Specifically, the fully connected layer of the CNN model calculates the probability of each type of carpet using the Softmax function, and selects the category with the highest probability as the final recognition result. Specifically, it outputs the probability of a small carpet, a large carpet, and no carpet. When the probability of a small carpet is the highest, the carpet recognition result is determined to be that the floor to be cleaned has a small carpet; when the probability of a large carpet is the highest, the carpet recognition result is determined to be that the floor to be cleaned has a large carpet; when the probability of no carpet is the highest, the carpet recognition result is determined to be that the floor to be cleaned has no carpet.

[0033] The carpet recognition model in this embodiment is trained through the following steps: An initial recognition model is constructed and a sample set is obtained, which includes multiple small carpet sample images, multiple large carpet sample images, and multiple ground sample images; the initial recognition model is trained based on the sample set to obtain the carpet recognition model.

[0034] In this embodiment, the carpet sample images are those labeled with small carpet tags from the sample set, the large carpet sample images are those labeled with large carpet tags from the sample set, and the ground sample images are those labeled with and without carpet tags from the sample set. Specifically, the sample images from the sample set are input into the constructed initial recognition model for training. The model parameters are optimized using the cross-entropy loss function and the Adam optimizer. Through multiple rounds of iterative training, the model parameters are continuously adjusted until the number of iterations reaches a preset iteration threshold or the model accuracy reaches a preset accuracy threshold.

[0035] In this embodiment, the sample set is split into a training sample set and a validation sample set. The training sample set is used to train the initial recognition model, and the validation sample set is used to validate the trained recognition model to obtain the model's accuracy.

[0036] The training and validation steps of the above model can also be performed in other devices, and then transferred to the cleaning robot after training is completed.

[0037] In some embodiments, the above sample set is obtained in the following manner: First, images of different types of carpets and carpetless floors are collected in various scenarios to obtain an initial dataset. Then, each image in the initial dataset is manually labeled, with labels for small carpet, large carpet, and no carpet added to the corresponding images to obtain multiple sample images. Finally, each sample image is enhanced by methods such as rotation, scaling, and flipping to increase the diversity of samples and improve the generalization ability of the model.

[0038] When annotating carpet images, annotators can classify carpets based on experience, for example, by their area (e.g., carpets smaller than an area threshold of 1.5m²). 2 Carpets can be classified in several ways: (1) carpets with an area of ​​less than or equal to a certain threshold are considered small carpets, while those with an area of ​​less than or equal to a certain threshold are considered large carpets; or carpets can be classified by a combination of their area and thickness; or by their width perpendicular to the direction the cleaning robot is moving (e.g., carpets with a width greater than 1m are considered large carpets, while those less than or equal to 1m are considered small carpets); or by their width perpendicular to the direction the cleaning robot is moving and their thickness.

[0039] Furthermore, to improve the model's recognition accuracy, in some embodiments, the carpet recognition model is trained online. In some embodiments, the online training of the model is specifically carried out through the following steps: During the operation of the cleaning robot, multiple initial images are acquired through the image acquisition device; the carpet types of the multiple initial images are labeled to obtain multiple online sample images; and the carpet recognition model is updated based on the multiple online sample images.

[0040] Specifically, after the cleaning robot collects multiple initial images, it can push the collected initial images to the user terminal (such as a mobile phone) bound to the cleaning robot. After the user views the initial images through the user terminal, the user can manually annotate the initial images and add corresponding carpet tags, including small carpet tags, large carpet tags, and no carpet tags, to obtain online sample images. Then, the carpet recognition model is further trained based on the online sample images to update the carpet recognition model.

[0041] As can be seen, in this embodiment, when updating the model, the collected images are pushed to the bound user terminal without sending them to external devices. Since the initial images collected are generally images of the user's home, updating the model through this embodiment can improve the privacy of user data. In addition, the online sample images in this embodiment are images of the area to be cleaned by the cleaning robot. Using the images of the area cleaned by the cleaning robot as samples to perform personalized training on the model in the cleaning robot can improve the model's recognition accuracy.

[0042] S130. When the carpet identification result indicates that a small carpet is laid on the floor to be cleaned, the cleaning mode of the cleaning robot is adjusted to carpet cleaning mode, and at least one small carpet cleaning path is generated according to the preset small carpet path planning algorithm.

[0043] In this embodiment, when the image of the floor to be cleaned identifies that a small carpet is laid on the floor to be cleaned, the cleaning robot's cleaning mode is immediately adjusted to carpet cleaning mode. The cleaning mode is adjusted to carpet cleaning mode in advance before the small carpet is laid, that is, the mop is lifted in advance to avoid the mop from contacting the small carpet, thereby reducing the probability that the small carpet will be dragged or rolled up by the cleaning robot. In addition, it also avoids the mop from getting the carpet dirty.

[0044] Specifically, when the cleaning robot's cleaning mode is adjusted to carpet cleaning mode, the cleaning robot's mop needs to be lifted, and the cleaning robot's working parameters need to be adjusted to the carpet working parameters corresponding to the carpet cleaning mode. The working parameters include suction parameters, mop rotation speed, and travel speed. The mop includes a mop plate or a roller brush.

[0045] Compared to the floor cleaning mode, the cleaning robot's mop needs to be lifted in the carpet cleaning mode. The suction parameters, mop rotation speed, and travel speed are all lower in the carpet cleaning mode to avoid rolling up the carpet while cleaning it.

[0046] Furthermore, in some embodiments, the cleaning robot is also equipped with a laser sensor, and the method further includes the following steps before adjusting the cleaning mode of the cleaning robot to carpet cleaning mode: The thickness of the small rug is obtained through the laser sensor; the rug cleaning parameters in the rug cleaning mode are adjusted according to the thickness.

[0047] Specifically, the thickness of the small carpet is obtained by emitting a vertical laser through a laser sensor. This thickness is used to determine the suction parameter in the carpet's working parameters. Within a preset carpet suction range, the thicker the small carpet, the greater the suction.

[0048] In some embodiments, a small carpet cleaning path is generated through the following steps: Obtain the width of the small carpet perpendicular to the direction of travel of the cleaning robot; generate at least one cleaning path for the small carpet based on the carpet width and the body width of the cleaning robot, wherein when the cleaning robot is in any of the small carpet cleaning paths, the edge of the cleaning robot's body is located within the small carpet. Specifically, the edge of the body is located within the carpet area corresponding to the longest width of the small carpet (the carpet width perpendicular to the direction of travel of the cleaning robot), so as to ensure that the cleaning robot is within the small carpet when cleaning the small carpet based on the small carpet cleaning path.

[0049] Specifically, after obtaining the carpet width, the edge carpet point is determined based on the preset gap distance and the body width of the cleaning robot. The distance between the position of the edge carpet point and the first side edge of the small carpet is the first distance, which is the sum of half the body width and the gap distance. Then, a small carpet cleaning path is generated based on the edge carpet point. Then, other small carpet cleaning paths are generated sequentially from the first side edge to the second side edge of the small carpet according to the preset cleaning interval, until the interval between the generated small carpet cleaning path and the second side edge of the small carpet is less than the second distance, which is the sum of the first distance and the preset cleaning interval.

[0050] The cleaning path of the small carpet is parallel to the direction of travel. When the cleaning robot is on the cleaning path of the small carpet, the center of the cleaning robot coincides with the cleaning path of the small carpet. The first side edge is the left or right edge of the small carpet (specifically, the side closest to the current direction of travel of the cleaning robot; if the left edge is closest, the first side edge is the left edge of the small carpet; otherwise, the first side edge is the right edge of the small carpet). When the first side edge is the left edge of the small carpet, the second side edge is the right edge of the small carpet. When the first side edge is the right edge of the small carpet, the second side edge is the left edge of the small carpet.

[0051] The preset gap distance is greater than or equal to 0cm, which is set by the user or specified in the cleaning robot cleaning system. It is the minimum distance between the edge of the cleaning robot body and the edge of the carpet when cleaning a small carpet.

[0052] The following is a detailed description of the process of generating a small carpet cleaning path using a small carpet path planning algorithm in this invention, using a specific example: like Figure 2 As shown, Figure 2This is a schematic diagram of a rectangular carpet path planning provided in an embodiment of the present invention. When the detected carpet width is 100cm (the longest distance of the carpet perpendicular to the direction of movement of the cleaning robot; if the rectangular carpet is placed perpendicular to the direction of movement of the cleaning robot, then the carpet width and carpet side length are included; otherwise, the carpet side length is not included), the cleaning robot body width is 32cm, the preset cleaning interval is 30cm, and the preset gap distance is 2cm, the first distance is 32 / 2+2=18cm, and the second distance is 18+30=48cm. Therefore, the generated first carpet cleaning path (No. 1) is 18cm from the first side edge of the carpet. At this point, the remaining unprocessed paths... The planned carpet width (i.e., the distance between the generated small carpet cleaning path and the second edge of the small carpet) is 100-17=83cm, which is greater than the second distance. Continue to generate the next path. The generated second small carpet cleaning path (No. 2) is 17+30=47cm from the first edge. At this time, the width of the remaining carpet without path planning is 100-47=53cm, which is greater than the second distance. Continue to generate the next path. The generated third small carpet cleaning path (No. 3) is 47+30=77cm from the first edge. At this time, the width of the remaining carpet without path planning is 100-77=23cm, which is less than the second distance. The small carpet cleaning path planning is completed.

[0053] Figure 2 This explanation uses a rectangular rug as an example, but some households use round or oval rugs. In this case, when generating the cleaning path for the rug, the obtained rug width is the longest width of the rug perpendicular to the direction of the cleaning robot's movement. For example, ... Figure 3 As shown, Figure 3 This is a schematic diagram of a circular carpet path planning provided in an embodiment of the present invention. If the carpet is circular, the carpet width is the diameter of the carpet. In this case, the two endpoints of the diameter perpendicular to the forward direction of the cleaning robot are respectively determined as the first side edge and the second side edge on the extension line of the forward direction of the cleaning robot.

[0054] Furthermore, if the currently laid carpet is detected to be a mini rug, since the area to be cleaned is too small, to avoid soiling the carpet on the roller track, the mini rug is not cleaned at this time. Instead, it is marked as an obstacle and avoided during cleaning. In this case, after obtaining the width of the mini rug perpendicular to the direction of travel of the cleaning robot, the method further includes: Determine whether the width of the carpet is greater than or equal to a preset threshold for the width of a small carpet (e.g., 35cm). At this point, the corresponding small carpet cleaning path is generated only after ensuring that the carpet width is greater than or equal to the small carpet width threshold. If the carpet width is less than the preset small carpet width threshold, the small carpet will not be cleaned. If it is less than the small carpet width threshold, the small carpet will be marked as an obstacle and the mini carpet will not be cleaned.

[0055] S140. Perform carpet cleaning work on the small carpet according to at least one of the small carpet cleaning paths.

[0056] In this embodiment, the cleaning path of each small carpet is used to guide the cleaning robot to the carpet mounting point and the direction of movement, and may not include the path length.

[0057] When there is only one cleaning path for the small rug, the rug is placed directly from the placement point indicated by the cleaning path, and the small rug is cleaned along the cleaning path until it is removed, thus completing the cleaning work.

[0058] When the small carpet cleaning path includes multiple paths, in some embodiments, step S140 includes: determining the current cleaning path from the multiple small carpet cleaning paths; placing the carpet on the current cleaning path and performing a cleaning operation on the small carpet according to the current cleaning path; when it is detected that the cleaning robot has removed the carpet based on the current cleaning path, selecting an unexecuted small carpet cleaning path from the multiple small carpet cleaning paths to update it as the current cleaning path, and returning to execute the steps of placing the carpet on the current cleaning path and performing a cleaning operation on the small carpet according to the current cleaning path, until all the multiple small carpet cleaning paths have been executed.

[0059] Specifically, multiple small carpet cleaning paths can carry a corresponding execution order, which can be the order in which the paths are generated. When cleaning the carpet, the small carpet cleaning paths can be executed sequentially according to the execution order of each path to clean the carpet. At this time, the cleaning robot needs to determine the current cleaning path and update the current cleaning path according to the execution order.

[0060] In addition, the generated small carpet cleaning path may not carry a corresponding execution order. In this case, the cleaning robot randomly determines the current cleaning path from the multiple generated small carpet cleaning paths and updates the current cleaning path randomly with the small carpet cleaning paths that have never been executed.

[0061] In this embodiment, after cleaning the small rug, the area where the rug is located needs to be marked as cleaned on the cleaning map of the current cleaning cycle. The cleaning robot is then moved to the position it was in before cleaning the rug and continues cleaning. Since the area where the rug is located is already cleaned, the cleaning robot needs to bypass this area when cleaning according to the room's predetermined cleaning path, and then continue cleaning according to the predetermined path. In some embodiments, the cleaning robot's detour trajectory is as follows: Figure 4 As shown, the target room (target cleaning area) is the room that the cleaning robot is currently cleaning, and the small carpet is laid in the target room. The predetermined cleaning path is the cleaning path set by the cleaning robot for cleaning the target room.

[0062] In summary, before cleaning the floor, this embodiment of the invention first determines whether a carpet is laid on the floor by acquiring an image of the floor to be cleaned. If a small carpet is laid, the cleaning mode is adjusted to carpet cleaning mode in advance to ensure that the mop is lifted in advance before the small carpet is placed on it, avoiding contact between the mop and the small carpet, thereby reducing the probability of the small carpet being dragged or rolled up by the cleaning robot. On the other hand, before cleaning the small carpet, this embodiment of the invention generates at least one small carpet cleaning path according to a preset small carpet path planning algorithm. There is no need to use a predetermined cleaning path for carpet cleaning. The cleaning robot can adjust its driving position through the generated small carpet cleaning path to avoid the robot's edge going beyond the edge of the small carpet, thereby further avoiding the situation where the small carpet is dragged or rolled up by the cleaning robot.

[0063] Figure 5 This is a schematic flowchart of a carpet cleaning method provided in another embodiment of the present invention. Figure 5 As shown, the carpet cleaning method of this embodiment includes steps S510-S560. Steps S510-S540 are similar to steps S110-S140 in the above embodiment and will not be described again here. The following details the additional steps S550-S570 in this embodiment.

[0064] S510. Obtain the target image of the ground to be cleaned through the image acquisition device.

[0065] S520. Perform carpet recognition processing on the ground to be cleaned according to the target image to obtain carpet recognition results. When the carpet recognition result indicates that the ground to be cleaned is covered with a small carpet, proceed to steps S530-S540; when the carpet recognition result indicates that the ground to be cleaned is covered with a large carpet, proceed to steps S550-S560; when the carpet recognition result indicates that the ground to be cleaned is not covered with a carpet, proceed to step S570.

[0066] S530. Adjust the cleaning mode of the cleaning robot to carpet cleaning mode, and generate at least one small carpet cleaning path according to the preset small carpet path planning algorithm.

[0067] S540. Perform carpet cleaning work on the small carpet according to at least one of the small carpet cleaning paths.

[0068] S550. Adjust the cleaning mode of the cleaning robot to carpet cleaning mode, and generate a carpet cleaning path according to the preset carpet path planning algorithm.

[0069] In this embodiment, when the carpet identification result indicates that the floor to be cleaned is covered with a large carpet, the cleaning robot's cleaning mode is immediately adjusted to carpet cleaning mode, and the mop is lifted before contacting the edge of the large carpet to avoid the mop getting dirty on the carpet.

[0070] In some embodiments, the carpet cleaning mode in this embodiment includes a large carpet cleaning mode and a small carpet cleaning mode. When a small carpet is detected on the floor to be cleaned, the cleaning robot's cleaning mode is adjusted to the small carpet cleaning mode. When a large carpet is detected on the floor to be cleaned, the cleaning robot's cleaning mode is adjusted to the large carpet cleaning mode. The travel speed in the large carpet cleaning mode is greater than the travel speed in the small carpet cleaning mode, so as to improve the cleaning efficiency of the large carpet.

[0071] In some embodiments, the present invention generates a carpet cleaning path through the following steps: moving the cleaning robot using a preset line-following algorithm to obtain boundary data of the carpet; and generating the carpet cleaning path based on the boundary data.

[0072] Specifically, the present invention moves the cleaning robot to perform carpet-walking operations using a line-following algorithm, and acquires a binarized image of the carpet edge using an image acquisition device. By extracting the linear features of the binarized image, the boundary data of the carpet is obtained. Finally, a carpet cleaning path is generated inside the carpet based on the boundary data. Specifically, the carpet cleaning path can be an arc-shaped cleaning path.

[0073] in, Figure 6 This is a schematic diagram of a cleaning robot performing a carpet-circling operation to obtain boundary data (carpet path) in an embodiment of the present invention.

[0074] Furthermore, users can also set the carpet cleaning depth. In this case, the cleaning robot needs to generate a carpet cleaning path based on the carpet path planning algorithm and the currently set carpet cleaning depth. The higher the carpet cleaning depth, the smaller the interval between adjacent paths in the generated carpet cleaning path.

[0075] To further understand the carpet cleaning path generated by this invention, please refer to [link / reference]. Figure 7 as well as Figure 8 ,in, Figure 7 This is a schematic diagram of the path planning for the cleaning path of a rectangular carpet. Figure 8 This is a schematic diagram of the cleaning path planning for a circular carpet.

[0076] S560. Perform carpet cleaning work on the carpet within the carpet according to the carpet cleaning path.

[0077] In this embodiment, after generating the carpet cleaning path, the cleaning robot puts on the carpet in cleaning mode and performs carpet cleaning work on the carpet based on the carpet cleaning path. Only after the carpet cleaning is completed through the carpet cleaning path is the carpet removed.

[0078] As can be seen, by performing the cleaning operation on a large carpet through the embodiments of the present invention, once the large carpet is identified, the cleaning of the large carpet can be completed in one step. The cleaning robot does not need to perform multiple carpet-mounting and unmounting operations, nor does it need to repeatedly identify the large carpet or repeatedly change the cleaning mode, thus improving the carpet cleaning efficiency.

[0079] In this embodiment, after the carpet is cleaned, the area where the carpet is located needs to be marked as cleaned on the cleaning map of the current cleaning cycle. The cleaning robot is then moved to the position it was in before cleaning the carpet and continues the floor cleaning work. Since the floor where the carpet is located is already cleaned, the cleaning robot needs to bypass this area when cleaning according to the room's predetermined cleaning path, and then continue cleaning according to the room's predetermined cleaning path.

[0080] Furthermore, after obtaining the carpet recognition result, the cleaning robot in this embodiment of the invention also updates the carpet identifier corresponding to the ground to be cleaned in the preset carpet map according to the carpet recognition result. The carpet map stores carpet information in the target cleaning area. The carpet information includes historical carpet tags obtained according to historical carpet recognition results and the ground location information corresponding to the historical carpet tags. The carpet map is used to assist the cleaning robot in performing carpet cleaning work.

[0081] If the carpet marker in a certain area is detected to be different from the historical marker based on the recognition results, the latest detected carpet marker will be updated to the carpet marker of that area, so as to update the carpet status of the carpet map to the latest status.

[0082] Because the cleaning robot of the present invention stores a carpet map, the cleaning robot can know in advance which locations in the target cleaning area are covered with carpet. At this time, before the cleaning robot arrives at the cleaning area covered with carpet, it can set the cleaning mode of the cleaning robot to carpet cleaning mode in advance according to the instructions of the carpet map. Furthermore, the carpet map can assist the cleaning robot in cleaning carpets placed in the blind spot of the cleaning robot.

[0083] S570. Perform floor cleaning work on the floor to be cleaned according to the preset floor cleaning mode.

[0084] In this embodiment, when the carpet identification result indicates that there is no carpet on the floor to be cleaned, it is necessary to further detect whether there are obstacles in the area. If there are no obstacles, the floor cleaning work is performed on the floor to be cleaned according to the predetermined cleaning path based on the floor cleaning mode. If there are obstacles, the obstacles are bypassed according to the preset obstacle avoidance strategy, and then the floor cleaning work is performed according to the predetermined cleaning path.

[0085] Figure 9 This is a schematic block diagram of a carpet cleaning device 900 provided in an embodiment of the present invention. Figure 9 As shown, corresponding to the above carpet cleaning methods, the present invention also provides a carpet cleaning device 900. This carpet cleaning device 900 includes a unit for performing the above-described carpet cleaning methods, and the device is configured in a cleaning robot. Specifically, please refer to... Figure 9 The carpet cleaning device 900 includes a receiving and dispatching unit 901 and a processing unit 902.

[0086] In some embodiments, when the small carpet cleaning path includes multiple paths, the processing unit 902, when performing the carpet cleaning step according to at least one of the small carpet cleaning paths, is specifically used for: The robot determines the current cleaning path from among multiple small carpet cleaning paths; it applies the carpet based on the current cleaning path and performs cleaning operations on the small carpet according to the current cleaning path; when it detects that the cleaning robot has removed the carpet based on the current cleaning path, it selects an unexecuted small carpet cleaning path from among the multiple small carpet cleaning paths to update it as the current cleaning path, and returns to execute the steps of applying the carpet based on the current cleaning path and performing cleaning operations on the small carpet according to the current cleaning path, until all multiple small carpet cleaning paths have been executed.

[0087] In some embodiments, when the processing unit 902 performs the step of generating at least one small carpet cleaning path on the small carpet according to a preset small carpet path planning algorithm, it is specifically used for: Obtain the width of the small carpet perpendicular to the direction of travel of the cleaning robot; At least one small carpet cleaning path is generated based on the carpet width and the body width of the cleaning robot, wherein when the cleaning robot is in any of the small carpet cleaning paths, the edge of the cleaning robot's body is located inside the small carpet.

[0088] In some embodiments, when performing the step of adjusting the cleaning mode of the cleaning robot to the carpet cleaning mode, the processing unit 902 is specifically used for: Lift the mop of the cleaning robot and adjust the working parameters of the cleaning robot to the carpet working parameters corresponding to the carpet cleaning mode. The working parameters include suction parameters, the rotation speed of the mop, and the travel speed. The mop includes a mop plate or a roller brush.

[0089] In some embodiments, after performing the carpet cleaning steps according to at least one of the carpet cleaning paths, the processing unit 902 is further configured to: Once the small rug is cleaned, the area where the small rug is located is marked as cleaned on the cleaning map for the current cleaning cycle.

[0090] In some embodiments, after performing the step of performing carpet recognition processing on the floor to be cleaned based on the target image to obtain carpet recognition results, the processing unit 902 is further configured to: When the carpet identification result indicates that the floor to be cleaned is covered with a large carpet, the cleaning robot's cleaning mode is adjusted to carpet cleaning mode, and a large carpet cleaning path is generated according to a preset large carpet path planning algorithm; the large carpet is then cleaned within the large carpet according to the large carpet cleaning path.

[0091] In some embodiments, when the processing unit 902 performs the step of generating a carpet cleaning path according to a preset carpet path planning algorithm, it is specifically used for: The cleaning robot moves using a preset line-following algorithm to obtain the boundary data of the large carpet; The carpet cleaning path is generated based on the boundary data.

[0092] In some embodiments, the cleaning robot is equipped with an ultrasonic sensor on its bottom, and before performing the step of acquiring the target image of the floor to be cleaned via the image acquisition device, the processing unit 902 is further configured to: The transceiver unit 901 obtains the ground material of the current ground through the ultrasonic sensor; when the ground material is carpet and the current cleaning mode is non-carpet cleaning mode, the cleaning robot is controlled to retreat a preset retreat distance so that the cleaning robot is at the image acquisition point of the ground to be cleaned, and a carpet detection command is generated. At this time, when the processing unit 902 performs the step of acquiring the target image of the ground to be cleaned through the image acquisition device, it is specifically used for: In response to the carpet detection command, the target image is acquired at the image acquisition point by the image acquisition device.

[0093] In some embodiments, the cleaning robot is further equipped with a laser sensor, and the processing unit 902 is further configured to: Before performing the step of adjusting the cleaning mode of the cleaning robot to carpet cleaning mode, The transceiver unit 901 obtains the thickness of the small carpet through the laser sensor; and adjusts the carpet working parameters in the carpet cleaning mode according to the thickness.

[0094] In some embodiments, after performing the step of performing carpet recognition processing on the floor to be cleaned based on the target image to obtain carpet recognition results, the processing unit 902 is further configured to: The carpet identification results are used to update the carpet identifiers corresponding to the floor to be cleaned in the preset carpet map. The carpet map stores carpet information in the target cleaning area. The carpet information includes historical carpet tags obtained from historical carpet identification results and the floor location information corresponding to the historical carpet tags. The carpet map is used to assist the cleaning robot in carpet cleaning.

[0095] In some embodiments, when the processing unit 902 performs the step of performing carpet recognition processing on the floor to be cleaned based on the target image to obtain carpet recognition results, it is specifically used for: The target image is preprocessed to obtain a processed image. The preprocessing includes size scaling and pixel normalization. The processed image is then input into a preset carpet recognition model for carpet recognition processing to obtain the carpet recognition result.

[0096] In some embodiments, before performing the step of inputting the processed image into a preset carpet recognition model for carpet recognition processing to obtain the carpet recognition result, the processing unit 902 is further configured to: An initial recognition model is constructed and a sample set is obtained, which includes multiple small carpet sample images, multiple large carpet sample images, and multiple ground sample images; the initial recognition model is trained based on the sample set to obtain the carpet recognition model.

[0097] In some embodiments, after performing the step of training the initial recognition model based on the sample set to obtain the carpet recognition model, the processing unit 902 is further configured to: During the operation of the cleaning robot, multiple initial images are acquired by the transceiver unit 901 through the image acquisition device; the carpet type of the multiple initial images is labeled to obtain multiple online sample images; and the carpet recognition model is updated based on the multiple online sample images.

[0098] In summary, when the carpet cleaning device 900 provided in this embodiment controls the cleaning robot, on the one hand, before cleaning the floor to be cleaned, it first determines whether a carpet is laid on the floor by acquiring an image of the floor to be cleaned. If a small carpet is laid, the cleaning mode is adjusted to carpet cleaning mode in advance to ensure that the mop is lifted in advance before the small carpet is placed on it, avoiding contact between the mop and the small carpet, thereby reducing the probability of the small carpet being dragged or rolled up by the cleaning robot. On the other hand, before cleaning the small carpet, this embodiment generates at least one small carpet cleaning path according to a preset small carpet path planning algorithm. There is no need to use a predetermined cleaning path for carpet cleaning. The cleaning robot can adjust its driving position through the generated small carpet cleaning path to avoid the robot's edge from going beyond the edge of the small carpet, thereby further avoiding the situation where the small carpet is dragged or rolled up by the cleaning robot.

[0099] The aforementioned carpet cleaning device can be implemented as a computer program, which can, for example... Figure 10 The cleaning robot shown is running.

[0100] Please see Figure 10 , Figure 10 This is a schematic block diagram of a cleaning robot provided in an embodiment of the present invention. The cleaning robot 1000 is equipped with an image acquisition device.

[0101] See Figure 10 The cleaning robot 1000 includes a processor 1002, a memory, and a network interface 1005 connected via a system bus 1001. The memory may include a non-volatile storage medium 1003 and internal memory 1004.

[0102] The non-volatile storage medium 1003 may store an operating system 10031 and a computer program 10032. The computer program 10032 includes program instructions that, when executed, cause the processor 1002 to perform a carpet cleaning method.

[0103] The processor 1002 provides computing and control capabilities to support the operation of the entire cleaning robot 1000.

[0104] The internal memory 1004 provides an environment for the operation of the computer program 10032 in the non-volatile storage medium 1003. When the computer program 10032 is executed by the processor 1002, the processor 1002 can perform a carpet cleaning method.

[0105] This network interface 1005 is used for network communication with other devices. Those skilled in the art will understand that... Figure 10 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the cleaning robot 1000 to which the present invention is applied. The specific cleaning robot 1000 may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0106] The processor 1002 is used to run a computer program 10032 stored in the memory to perform the following steps: The target image of the ground to be cleaned is acquired using the image acquisition device; Based on the target image, carpet recognition processing is performed on the floor to be cleaned to obtain carpet recognition results; When the carpet identification result indicates that a small carpet is laid on the floor to be cleaned, the cleaning robot's cleaning mode is adjusted to carpet cleaning mode, and at least one small carpet cleaning path is generated according to the preset small carpet path planning algorithm. Carpet cleaning is performed on the small carpet according to at least one of the small carpet cleaning paths.

[0107] It should be understood that, in this embodiment of the invention, the processor 1002 may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0108] It will be understood by those skilled in the art that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program includes program instructions and can be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the process steps of the embodiments of the above methods.

[0109] Therefore, the present invention also provides a storage medium. This storage medium can be a computer-readable storage medium. The storage medium stores a computer program, wherein the computer program includes program instructions. When executed by a processor, the program instructions cause the processor to perform the following steps: The target image of the ground to be cleaned is acquired using the image acquisition device; Based on the target image, carpet recognition processing is performed on the floor to be cleaned to obtain carpet recognition results; When the carpet identification result indicates that a small carpet is laid on the floor to be cleaned, the cleaning robot's cleaning mode is adjusted to carpet cleaning mode, and at least one small carpet cleaning path is generated according to the preset small carpet path planning algorithm. Carpet cleaning is performed on the small carpet according to at least one of the small carpet cleaning paths.

[0110] The storage medium can be any computer-readable storage medium capable of storing program code, such as a USB flash drive, portable hard drive, read-only memory (ROM), magnetic disk, or optical disk.

[0111] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0112] In the several embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For example, the division of each unit is merely a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0113] The steps in the method of this invention can be adjusted, merged, or reduced in order according to actual needs. The units in the device of this invention can be merged, divided, or reduced according to actual needs. Furthermore, the functional units in the various embodiments of this invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0114] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a terminal, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.

[0115] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A carpet cleaning method, characterized in that: Applied to a cleaning robot, the cleaning robot is equipped with an image acquisition device, and the carpet cleaning method comprises: Acquire a target image of the ground to be cleaned by the image acquisition device; Performing carpet recognition processing on the floor to be cleaned according to the target image to obtain a carpet recognition result; When the carpet identification result indicates that the floor to be cleaned is paved with a small carpet, the cleaning mode of the cleaning robot is adjusted to a carpet cleaning mode, and at least one small carpet cleaning path is generated according to a preset small carpet path planning algorithm; A carpet cleaning operation is performed on the small carpet according to at least one of the small carpet cleaning paths.

2. The carpet cleaning method according to claim 1, characterized in that: When the carpet cleaning paths include a plurality of paths, performing carpet cleaning on the carpet according to at least one of the carpet cleaning paths includes: determining a current cleaning path from among the plurality of said carpet cleaning paths; placing a carpet based on the current cleaning path, and performing a cleaning operation on the carpet according to the current cleaning path; When it is detected that the cleaning robot is putting down the carpet based on the current cleaning path, an unexecuted small carpet cleaning path is selected from the multiple small carpet cleaning paths and updated as the current cleaning path, and the process returns to execute the steps of putting down the carpet based on the current cleaning path and cleaning the small carpet according to the current cleaning path until all the multiple small carpet cleaning paths are executed.

3. The carpet cleaning method according to claim 1, characterized in that: The step of generating at least one carpet cleaning path on the carpet according to a preset carpet path planning algorithm comprises: Obtaining the width of the small carpet perpendicular to the forward direction of the cleaning robot; At least one of the small carpet cleaning paths is generated according to the carpet width and the body width of the cleaning robot, wherein when the cleaning robot is in any of the small carpet cleaning paths, the body edge of the cleaning robot is located within the small carpet.

4. The carpet cleaning method according to claim 1, characterized in that: After performing carpet cleaning on the small carpet according to at least one of the small carpet cleaning paths, the method further includes: When the carpet is cleaned, the floor area where the carpet is located is marked as a cleaned floor in the cleaning map of the current cleaning cycle.

5. The carpet cleaning method according to claim 1, characterized in that: After performing carpet recognition processing on the floor to be cleaned according to the target image and obtaining a carpet recognition result, the method further includes: When the carpet recognition result indicates that the floor to be cleaned is paved with a large carpet, the cleaning mode of the cleaning robot is adjusted to a carpet cleaning mode, and a large carpet cleaning path is generated according to a preset large carpet path planning algorithm; A carpet cleaning operation is performed on the large carpet within the large carpet according to the large carpet cleaning path.

6. The carpet cleaning method according to claim 5, characterized in that: The step of generating a large carpet cleaning path according to a preset large carpet path planning algorithm includes: Moving the cleaning robot by a preset line-finding algorithm to obtain boundary data of the large carpet; The large carpet cleaning path is generated according to the boundary data.

7. The carpet cleaning method according to claim 1, characterized in that: An ultrasonic sensor is installed at the bottom of the cleaning robot. Before acquiring the target image of the floor to be cleaned by the image acquisition device, the method further includes: Acquire the ground material of the current ground through the ultrasonic sensor; When the floor material is carpet material and the current cleaning mode is non-carpet cleaning mode, the cleaning robot is controlled to retreat a preset retreat distance so that the cleaning robot is at an image acquisition point of the floor to be cleaned, and a carpet detection instruction is generated; The step of acquiring a target image of the ground to be cleaned by the image acquisition device includes: In response to the carpet detection instruction, the target image is acquired by the image acquisition device at the image acquisition point.

8. The carpet cleaning method according to claim 1, wherein the cleaning robot is further equipped with a laser sensor, and before adjusting the cleaning mode of the cleaning robot to the carpet cleaning mode, the method further comprises: Obtaining the thickness of the small carpet by means of the laser sensor; The carpet operating parameters in the carpet cleaning mode are adjusted according to the thickness.

9. The carpet cleaning method according to any one of claims 1 to 8, wherein after performing carpet recognition processing on the floor to be cleaned according to the target image and obtaining a carpet recognition result, the method further comprises: The carpet identification corresponding to the ground to be cleaned in the preset carpet map is updated according to the carpet recognition result. The carpet map stores carpet information in the target cleaning area. The carpet information includes historical carpet labels obtained according to historical carpet recognition results and ground position information corresponding to the historical carpet labels. The carpet map is used to assist the cleaning robot in performing carpet cleaning work.

10. A carpet cleaning device, characterized in that: The carpet cleaning device is deployed in a cleaning robot, the cleaning robot is equipped with an image acquisition device, and the carpet cleaning device includes: A transceiver unit, used to obtain a target image of the ground to be cleaned through the image acquisition device; The processing unit is used to perform carpet recognition processing on the floor to be cleaned according to the target image to obtain a carpet recognition result; when the carpet recognition result shows that the floor to be cleaned is covered with a small carpet, the cleaning mode of the cleaning robot is adjusted to a carpet cleaning mode, and at least one small carpet cleaning path is generated according to a preset small carpet path planning algorithm; and carpet cleaning work is performed on the small carpet according to at least one of the small carpet cleaning paths.

11. A cleaning robot, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: The cleaning robot is equipped with an image acquisition device, and the processor implements the carpet cleaning method according to any one of claims 1 to 9 when executing the computer program.

12. A storage medium, characterized in that: The storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by a processor, the processor executes the carpet cleaning method according to any one of claims 1 to 9.

Images