A path planning method and device for a cleaning robot

Abstract

The present invention relates to the technical field of data processing, and discloses a path planning method and device for a cleaning robot; the present invention uses a smart phone to collect images and extract image features of a preset boundary, and the extracted preset boundary feature distribution will be transmitted to the cleaning robot. The cleaning robot will use a camera device to collect images and extract image features of surrounding scenes, and compare the extracted real-time image feature distribution with the preset boundary feature distribution, and judge the existence of the preset boundary according to the comparison result. The cleaning robot will build basic environment modeling, preset boundaries and areas to be cleaned according to positioning information, real-time image feature distribution and comparison results, and perform path planning on this basis to achieve overall cleaning of the cleaning area, thereby solving the problem of high cost of setting the working boundary of the cleaning robot by setting a laser device in the prior art.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a path planning method and device for a cleaning robot. Background Art

[0002] A cleaning robot is a robot that can automatically clean the floor, saving manpower and improving efficiency.

[0003] When using a cleaning robot at home, it is generally necessary to close the doors leading to the kitchen, bathroom, balcony, outdoor area and other areas to prevent the cleaning robot from entering these areas. Therefore, these areas are dirtier and greasy than indoors. Working in these areas can easily shorten the life of the cleaning robot. Users occasionally forget to close the door, and the door closing operation is relatively troublesome.

[0004] In the traditional design concept, the solution is to set up a laser device at the passage leading to these areas. When the cleaning robot detects the laser signal emitted by the laser device, the cleaning robot will stop going to these areas, thereby replacing the door closing operation.

[0005] However, the above solution requires the installation of a laser device, which is relatively expensive. In addition, when the cleaning robot is moved to another location to work, the laser device needs to be reset, which is relatively complicated and requires additional costs. Summary of the invention

[0006] The purpose of the present invention is to provide a path planning method and device for a cleaning robot, aiming to solve the problem of high cost in setting the working boundary of the cleaning robot by setting a laser device in the prior art.

[0007] The present invention is implemented in this way. In a first aspect, the present invention provides a path planning method for a cleaning robot, comprising:

[0008] Receiving a plurality of preset boundary images acquired by a smart phone connected to a signal of the cleaning robot, performing feature extraction based on the plurality of preset boundary images, and acquiring a preset boundary feature distribution;

[0009] Driving a motion device disposed on the cleaning robot to enable the cleaning robot to move;

[0010] Receiving a real-time image acquired by a camera device disposed on the cleaning robot, performing feature extraction based on the real-time image, and acquiring a feature distribution of the real-time image;

[0011] Receiving positioning information acquired by a positioning device provided on the cleaning robot;

[0012] Building basic environment modeling and the area to be cleaned based on the real-time image feature distribution and the positioning information, and performing path planning according to the area to be cleaned;

[0013] A comparison analysis is performed based on the real-time image feature distribution and the preset boundary feature distribution, and a virtual obstacle is set based on the basic environment modeling according to the result of the comparison analysis.

[0014] Preferably, the step of acquiring the preset boundary image includes:

[0015] The image of the area where the preset boundary is located is acquired through a smartphone;

[0016] The preset boundary is marked with lines on the acquired image to generate the preset boundary image.

[0017] Preferably, the step of extracting features based on the plurality of preset boundary images to obtain the preset boundary feature distribution includes:

[0018] Determine the preset boundary according to the line mark in the preset boundary image;

[0019] Extracting image features from the outsides of two endpoints of the preset boundary in the preset boundary image to obtain a first preset boundary feature;

[0020] Extracting image features on both inner and outer sides of the preset boundary in the preset boundary image to obtain a second preset boundary feature;

[0021] Clearing the line mark, and performing image feature extraction along the preset boundary in the preset boundary image to obtain a third preset boundary feature;

[0022] The first preset boundary feature, the second preset boundary feature and the third preset boundary feature are combined to generate the preset boundary feature distribution.

[0023] Preferably, the step of performing feature extraction based on the real-time image to obtain the feature distribution of the real-time image includes:

[0024] Analyzing the real-time image to obtain ground information and obstacle information in the real-time image;

[0025] Performing obstacle recognition and obstacle distance analysis on the obstacle information in the real-time image to generate a first real-time image feature;

[0026] Performing lateral and longitudinal change analysis on the ground information in the real-time image to generate a second real-time image feature;

[0027] Performing line segment area analysis on the ground information in the real-time image to generate a third real-time image feature;

[0028] The ground information, the obstacle information, the first real-time image feature, the second real-time image feature, and the third real-time image feature are combined to generate a real-time image feature distribution.

[0029] Preferably, the steps of constructing basic environment modeling and the area to be cleaned based on the real-time image feature distribution and the positioning information, and performing path planning according to the area to be cleaned include:

[0030] Generating relative position relationship information between the feature distributions of the real-time image according to the acquisition time of the real-time image and the acquisition time of the positioning information;

[0031] Combining the ground information and the obstacle information in the real-time image feature distribution with each other according to the relative position relationship information to generate the basic environment modeling;

[0032] Analyzing the basic environment modeling to generate an operable area of ​​the cleaning robot;

[0033] Dividing the operable area into a cleaned area and an area to be cleaned according to the positioning information;

[0034] Path planning is performed according to the positional relationship between the areas to be cleaned.

[0035] Preferably, the step of performing a comparison analysis based on the real-time image feature distribution and the preset boundary feature distribution, and setting a virtual obstacle based on the basic environment modeling according to the result of the comparison analysis includes:

[0036] Comparing and analyzing the real-time image feature distribution with the preset boundary feature distribution to determine whether the preset boundary exists in the real-time image;

[0037] According to the result of the determination, the real-time image with the preset boundary is marked as a virtual obstacle image, and the preset boundary in the virtual obstacle image is converted from the ground information to the obstacle information.

[0038] In a second aspect, the present invention provides a path planning device for a cleaning robot, comprising:

[0039] A preset unit is used to receive a plurality of preset boundary images acquired by a smart phone connected to a signal of the cleaning robot, perform feature extraction based on the plurality of preset boundary images, and acquire a preset boundary feature distribution;

[0040] A driving unit, used to drive a motion device provided on the cleaning robot to enable the cleaning robot to move;

[0041] A collection unit, used for receiving a real-time image acquired by a camera device arranged on the cleaning robot, performing feature extraction based on the real-time image, and acquiring a feature distribution of the real-time image;

[0042] A positioning unit, used to receive positioning information acquired by a positioning device provided on the cleaning robot;

[0043] A planning unit, which constructs basic environment modeling and an area to be cleaned based on the real-time image feature distribution and the positioning information, and performs path planning according to the area to be cleaned;

[0044] The analysis unit is used to compare and analyze the real-time image feature distribution with the preset boundary feature distribution, and to set virtual obstacles based on the basic environment modeling according to the result of the comparison and analysis.

[0045] The present invention provides a path planning method and device for a cleaning robot, which has the following beneficial effects:

[0046] The present invention uses a smart phone to capture images and extract image features of preset boundaries. The extracted preset boundary feature distribution will be transmitted to the cleaning robot. The cleaning robot will use a camera device to capture images and extract image features of surrounding scenes, and compare the extracted real-time image feature distribution with the preset boundary feature distribution. The existence of the preset boundary is determined according to the comparison result. The cleaning robot will build basic environment modeling, preset boundaries and areas to be cleaned according to the positioning information, real-time image feature distribution and comparison results, and perform path planning on this basis to achieve overall cleaning of the cleaning area, thereby solving the problem of high cost of setting the working boundary of the cleaning robot by setting a laser device in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a schematic diagram of the steps of a path planning method for a cleaning robot provided by an embodiment of the present invention;

[0048] Figure 2 It is a structural schematic diagram of a path planning device for a cleaning robot provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0050] The same or similar numbers in the drawings of this embodiment correspond to the same or similar parts; in the description of the present invention, it should be understood that if the terms "upper", "lower", "left", "right" and the like indicate directions or positional relationships based on the directions or positional relationships shown in the drawings, it is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction. Therefore, the terms describing the positional relationship in the drawings are only used for illustrative purposes and cannot be understood as limitations on this patent. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0051] The implementation of the present invention is described in detail below in conjunction with specific embodiments.

[0052] Reference Figure 1 , Figure 2 As shown, a preferred embodiment of the present invention is provided.

[0053] In a first aspect, the present invention provides a path planning method for a cleaning robot, comprising:

[0054] S1: Receive a number of preset boundary images obtained by a smart phone connected to the signal of the cleaning robot, perform feature extraction based on the preset boundary images, and obtain a preset boundary feature distribution.

[0055] S2: driving the motion device provided on the cleaning robot to make the cleaning robot move.

[0056] S3: Receive a real-time image acquired by a camera device installed on the cleaning robot, perform feature extraction based on the real-time image, and acquire feature distribution of the real-time image.

[0057] S4: Receive positioning information acquired by a positioning device installed on the cleaning robot.

[0058] S5: constructing basic environment modeling and the area to be cleaned based on the real-time image feature distribution and the positioning information, and performing path planning according to the area to be cleaned.

[0059] S6: performing a comparison analysis based on the real-time image feature distribution and the preset boundary feature distribution, and setting a virtual obstacle based on the basic environment modeling according to the result of the comparison analysis.

[0060] Specifically, an APP for remotely controlling the cleaning robot is downloaded on the smartphone, and the smartphone and the cleaning robot are connected via wireless signals, that is, the cleaning robot can be controlled via the smartphone.

[0061] More specifically, the key points of the present invention are to use a smart phone to photograph the area where the preset boundary is located, and transmit the image to the cleaning robot. The cleaning robot will extract features from the image, and when the cleaning robot is running, it will extract features from the surrounding scene. When the preset boundary is identified through feature comparison, the cleaning robot will stop and turn before the preset boundary. In other words, the cleaning range of the cleaning robot will not exceed the preset boundary.

[0062] More specifically, in traditional designs, it is usually necessary to set a laser emitter at the position of a preset boundary, and the preset boundary is formed by laser. A laser sensor is provided on the cleaning robot. When the laser sensor detects the preset boundary formed by the laser, it will stop and turn, thereby achieving control of the cleaning range of the cleaning robot.

[0063] It can be understood that, compared with the traditional design, in the method provided by the present invention, there is no need to set a laser sensor at the position of the preset boundary, which saves costs, and the method of setting the preset boundary is more efficient.

[0064] More specifically, in order to enable the cleaning robot to have the function of capturing images of the surrounding environment, a camera device needs to be installed on the cleaning robot. It should be noted that the camera device here can be a camera fixed on the cleaning robot, or it can be a mounting bracket for fixing a smartphone. By fixing the smartphone on the cleaning robot, the image capture function of the smartphone itself can be used to capture surrounding scenes. The advantage of this method is that there is no need to set up an additional camera, which saves manufacturing costs.

[0065] It should be noted that the above content is used to describe how to set and execute preset boundaries. In addition to the preset boundaries, there are also naturally existing boundaries in the cleaning area, that is, various obstacles in the cleaning area, such as walls, steps, etc.

[0066] It can be understood that the preset boundary and the natural boundary together constitute the cleaning range of the cleaning robot, and the work of the cleaning robot is to clean within this cleaning range. Specifically, the cleaning device of the cleaning robot is arranged at its bottom. That is to say, when the cleaning robot runs through a certain area, the area will be cleaned by the cleaning device. Therefore, in order to complete the cleaning of the cleaning area, the cleaning robot needs to run through all areas in the cleaning area.

[0067] Specifically, in order to enable the cleaning robot to run through all areas within the cleaning area and avoid the cleaning robot missing some areas within the cleaning area during operation, the method selected by the present invention is to obtain real-time image feature distribution and positioning information as the cleaning robot runs, and construct basic environment modeling according to the real-time image feature distribution and positioning information. The basic environment modeling depicts the cleaning area through 3D modeling. The basic environment modeling includes two types of modeling: ground and obstacle. Among them, the ground part is the area to be cleaned by the cleaning robot, that is, the area where the cleaning robot needs to operate.

[0068] More specifically, basic environment modeling is used to provide reference information for the path planning of the cleaning robot. Therefore, basic environment modeling does not need to completely restore the scenes in the cleaning area, but selectively restores the scenes in the cleaning area that can affect the cleaning robot.

[0069] For example: the height of the cleaning robot is 10CM. If the bottom height of a table top is 20CM, there is no need to restore the bottom of the table top, and only the table legs need to be modeled.

[0070] More specifically, when the cleaning robot recognizes a preset boundary, it will model the preset boundary and set virtual obstacles for the preset boundary in the basic environment modeling. At this time, the preset boundary has the same properties as the natural boundary. When the cleaning robot approaches the preset boundary, it will stop moving and turn in other directions just like approaching a natural boundary.

[0071] More specifically, the area to be cleaned is the area that the cleaning robot needs to clean, that is, the area that the cleaning robot needs to pass through, and the technical solution to achieve this function belongs to the conventional technology in cleaning robots.

[0072] The present invention provides a path planning method and device for a cleaning robot, which has the following beneficial effects:

[0073] The present invention uses a smart phone to capture images and extract image features of preset boundaries. The extracted preset boundary feature distribution will be transmitted to the cleaning robot. The cleaning robot will use a camera device to capture images and extract image features of surrounding scenes, and compare the extracted real-time image feature distribution with the preset boundary feature distribution. The existence of the preset boundary is determined according to the comparison result. The cleaning robot will build basic environment modeling, preset boundaries and areas to be cleaned according to the positioning information, real-time image feature distribution and comparison results, and perform path planning on this basis to achieve overall cleaning of the cleaning area, thereby solving the problem of high cost of setting the working boundary of the cleaning robot by setting a laser device in the prior art.

[0074] Preferably, the step of obtaining the preset boundary image includes:

[0075] S11: Acquire an image of the area where the preset boundary is located through a smart phone.

[0076] S12: Marking the preset boundary with lines on the acquired image to generate the preset boundary image.

[0077] Specifically, when the user sets a preset boundary for the cleaning robot, the preset boundary needs to be photographed through a smart phone that is signal-connected to the cleaning robot and is loaded with an APP for controlling the cleaning robot.

[0078] More specifically, the image captured of the preset boundary needs to include the preset boundary in its entirety and needs to include the scene around the preset boundary, so as to extract features of the preset boundary and generate a preset boundary feature distribution.

[0079] More specifically, the preset boundaries may include a variety of situations, for example, when the cleaning robot is used within a home, the kitchen and bathroom may be set as preset boundaries, and when the cleaning robot is used in a hotel, the door may be set as a set boundary.

[0080] It should be noted that, since the image includes the preset boundary and the scene around the preset boundary, the preset boundary needs to be determined in the image, and a specific determination method is to use line marks in the image to determine the preset boundary.

[0081] Preferably, the step of extracting features based on the plurality of preset boundary images to obtain the preset boundary feature distribution includes:

[0082] S13: Determine the preset boundary according to the line mark in the preset boundary image.

[0083] S14: extracting image features from the outsides of two end points of the preset boundary in the preset boundary image to obtain a first preset boundary feature.

[0084] S15: extracting image features on both inner and outer sides of the preset boundary in the preset boundary image to obtain a second preset boundary feature.

[0085] S16: Clear the line mark, and extract image features along the preset boundary in the preset boundary image to obtain a third preset boundary feature.

[0086] S17: combining the first preset boundary feature, the second preset boundary feature and the third preset boundary feature to generate the preset boundary feature distribution.

[0087] Specifically, after determining the preset boundary, it is necessary to extract features of the preset boundary. In the subsequent operation of the cleaning robot, the cleaning robot will also perform real-time image acquisition and feature extraction of the surrounding scene. When the consistency between the real-time image feature distribution and the preset boundary feature distribution reaches the preset standard, the cleaning robot will recognize the preset boundary.

[0088] More specifically, the characteristics of the preset boundary include the following parts:

[0089] The first preset boundary feature comes from the scene outside the two endpoints of the preset boundary. It can be understood that the preset boundary is usually in the shape of a line segment, and the line segment has two endpoints. Under normal circumstances, the line segment part of the preset boundary is the ground, and the two ends are obstacles. Obviously, compared with the line segment part, the two endpoints are more recognizable. Therefore, it is necessary to extract features of the scenes at both ends of the preset boundary as the first preset boundary feature.

[0090] The second preset boundary feature comes from the scenes on both sides of the preset boundary. It can be understood that the preset boundary is usually the dividing point between two different areas, such as the boundary between the living room and the kitchen. In home decoration, the floors of the living room and the kitchen are different, that is to say, there are differences on both sides of the preset boundary that separates the two. This difference enables the cleaning robot to recognize the preset boundary. Therefore, it is necessary to extract features from the inner and outer scenes of the preset boundary as the second preset boundary feature.

[0091] The third preset boundary feature comes from the preset boundary itself. It can be understood that there are differences between the inner and outer sides of the preset boundary. In some cases, the preset boundary itself has special identification points. For example, at the boundary position of different areas, there is a narrow area that is different from the areas on both sides. Therefore, it is necessary to clear the line markings and perform image feature extraction along the preset boundary in the image with the preset boundary to obtain the third preset boundary feature.

[0092] More specifically, the first preset boundary feature, the second preset boundary feature, and the third preset boundary feature are combined to generate a preset boundary feature distribution.

[0093] Preferably, the step of performing feature extraction based on the real-time image to obtain the feature distribution of the real-time image includes:

[0094] S31: Analyze the real-time image to obtain ground information and obstacle information in the real-time image.

[0095] S32: performing obstacle recognition and obstacle distance analysis on the obstacle information in the real-time image to generate a first real-time image feature.

[0096] S33: Performing lateral and longitudinal change analysis on the ground information in the real-time image to generate a second real-time image feature.

[0097] S34: Perform line segment area analysis on the ground information in the real-time image to generate a third real-time image feature.

[0098] S35: combining the ground information, the obstacle information, the first real-time image feature, the second real-time image feature, and the third real-time image feature to generate a real-time image feature distribution.

[0099] Specifically, there are two main uses for real-time image feature distribution. The first is to compare with the preset boundary feature distribution to determine the preset boundary. The second is to identify ground information and obstacle information to build basic environment modeling.

[0100] More specifically, for the first purpose, various forms of feature extraction are required for the real-time image. More specifically, the feature extraction of the real-time image can refer to various feature forms in the preset boundary feature distribution. That is, various features in the real-time image feature distribution need to correspond one-to-one with the preset boundary feature distribution.

[0101] More specifically, the real-time image feature distribution mainly includes the following parts:

[0102] The first real-time image feature is to perform obstacle recognition and obstacle distance analysis based on the obstacle information in the real-time image. It should be noted that the scenes at both ends of the preset boundary usually correspond to obstacles. Therefore, by identifying the obstacles and analyzing the obstacle distance, it can be determined whether they are the two end points of the preset boundary, thereby determining whether it is the preset boundary.

[0103] The second real-time image feature performs lateral and longitudinal change analysis based on the ground information in the real-time image. It should be noted that there are usually differences in ground scenes on both sides of the preset boundary. Therefore, through the changes in ground information, it can be identified whether the ground information comes from two different areas, and whether these two areas are consistent with the areas on both sides of the preset boundary, thereby determining the preset boundary.

[0104] The third real-time image feature is to analyze the line segment type area according to the ground information in the real-time image. It should be noted that the preset boundary itself is usually a line segment on the ground. Therefore, the line segment type area can be retrieved and analyzed on the ground information, thereby realizing the judgment of the preset boundary.

[0105] It should be noted that the above-mentioned first real-time image feature, second real-time image feature and third real-time image feature can all be used to compare with the preset boundary feature distribution, that is, the first real-time image feature, the second real-time image feature and the third real-time image feature are combined to generate a real-time image feature distribution, which can be used to compare with the preset boundary feature distribution to determine the preset boundary.

[0106] More specifically, when comparing the real-time image feature distribution with the preset boundary feature distribution, it is necessary to set an evaluation standard for the judgment program in advance. The purpose of the evaluation standard is to judge the scores obtained by comparing the various types of features of the real-time image feature distribution with the preset boundary feature distribution. When the total score exceeds the preset evaluation standard, it means that there is a preset boundary in the real-time image.

[0107] More specifically, after the preset boundary is identified, the distance between the cleaning robot and the preset boundary can be determined based on the position of the preset boundary in the real-time image, thereby controlling the cleaning robot to not run beyond the preset boundary.

[0108] Specifically, for the second purpose, the real-time image needs to be analyzed to obtain ground information and obstacle information in the real-time image.

[0109] It should be noted that the ground information represents the information of the ground, and the obstacle information represents the information of the obstacles. More specifically, the horizontal height that is consistent with the horizontal height of the plane where the cleaning robot is located is regarded as the ground, and the horizontal height that is higher or lower is regarded as an obstacle. In other words, the ground is the area that the cleaning robot can go to, and the obstacle is the area that prevents the operation of the cleaning robot.

[0110] Preferably, the steps of constructing basic environment modeling and the area to be cleaned based on the real-time image feature distribution and the positioning information, and performing path planning according to the area to be cleaned include:

[0111] S51: Generate relative position relationship information between the real-time image feature distributions according to the acquisition time of the real-time image and the acquisition time of the positioning information.

[0112] S52: combining the ground information and the obstacle information in the real-time image feature distribution according to the relative position relationship information to generate the basic environment modeling.

[0113] S53: Analyze the basic environment modeling to generate an operable area for the cleaning robot.

[0114] S54: Divide the operable area into a cleaned area and an area to be cleaned according to the positioning information.

[0115] S55: performing path planning according to the positional relationship between the areas to be cleaned.

[0116] Specifically, when the cleaning robot acquires a real-time image, the time of acquisition is determined, and the positioning information of the cleaning robot when acquiring the image is also determined. Therefore, it can be calculated that the relative position of the cleaning robot changes when the cleaning robot acquires two adjacent real-time images.

[0117] More specifically, the relative position relationship characteristics between the real-time image features can be obtained according to the relative position change, and the relative position relationship characteristics are used to describe the relative change of the observation angle of the ground information and the obstacle information in the real-time image.

[0118] More specifically, at time point A1, the positioning information of the cleaning robot is A2, and the cleaning robot obtains real-time image A3. At time point B1, the positioning information of the cleaning robot is B2, and the cleaning robot obtains real-time image B3.

[0119] More specifically, real-time image A3 and real-time image B3 contain ground information and obstacle information, respectively. Since the positioning information A2 of the cleaning robot at time point A1 is different from the positioning information B2 of the cleaning robot at time point B1, that is, the viewing angles of the cleaning robot when obtaining real-time image A3 and real-time image B3 are different, there are differences in the images obtained.

[0120] It is understandable that in order to construct basic environmental modeling, the ground information and obstacle information in the real-time image need to be converted to obtain basic environmental modeling. The basic environmental modeling constructed by a single real-time image is limited. In order to construct a complete basic environmental modeling, the partial environmental modeling constructed by each real-time image needs to be superimposed. How to superimpose requires reference to the relative position relationship information between the feature distributions of each real-time image.

[0121] Preferably, the step of performing a comparison analysis based on the real-time image feature distribution and the preset boundary feature distribution, and setting a virtual obstacle based on the basic environment modeling according to the result of the comparison analysis includes:

[0122] S61: performing a comparison analysis based on the real-time image feature distribution and the preset boundary feature distribution to determine whether the preset boundary exists in the real-time image.

[0123] S62: marking the real-time image with the preset boundary as a virtual obstacle image according to the result of the determination, and converting the preset boundary in the virtual obstacle image from the ground information to the obstacle information.

[0124] Specifically, under normal circumstances, the preset boundary belongs to ground information. In order to generate a virtual obstacle in basic environment modeling according to the preset boundary, the preset boundary needs to be converted from ground information to obstacle information, and then processed according to the obstacle information to generate a virtual obstacle.

[0125] In a second aspect, the present invention provides a path planning device for a cleaning robot, comprising:

[0126] A preset unit is used to receive a plurality of preset boundary images acquired by a smart phone connected to a signal of the cleaning robot, perform feature extraction based on the plurality of preset boundary images, and acquire a preset boundary feature distribution;

[0127] A driving unit, used to drive a motion device provided on the cleaning robot to enable the cleaning robot to move;

[0128] A collection unit, used for receiving a real-time image acquired by a camera device arranged on the cleaning robot, performing feature extraction based on the real-time image, and acquiring a feature distribution of the real-time image;

[0129] A positioning unit, used to receive positioning information acquired by a positioning device provided on the cleaning robot;

[0130] A planning unit, which constructs basic environment modeling and an area to be cleaned based on the real-time image feature distribution and the positioning information, and performs path planning according to the area to be cleaned;

[0131] The analysis unit is used to compare and analyze the real-time image feature distribution with the preset boundary feature distribution, and to set virtual obstacles based on the basic environment modeling according to the result of the comparison and analysis.

[0132] In this embodiment, for the specific implementation of each unit in the above device embodiment, please refer to the above method embodiment, which will not be repeated here.

[0133] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A path planning method for a cleaning robot, characterized in that: include: Receiving a plurality of preset boundary images acquired by a smart phone connected to a signal of the cleaning robot, performing feature extraction based on the plurality of preset boundary images, and acquiring a preset boundary feature distribution; Driving a motion device disposed on the cleaning robot to enable the cleaning robot to move; Receiving a real-time image acquired by a camera device disposed on the cleaning robot, performing feature extraction based on the real-time image, and acquiring a feature distribution of the real-time image; Receiving positioning information acquired by a positioning device provided on the cleaning robot; Building basic environment modeling and the area to be cleaned based on the real-time image feature distribution and the positioning information, and performing path planning according to the area to be cleaned; Comparing and analyzing the real-time image feature distribution with the preset boundary feature distribution, and setting a virtual obstacle based on the basic environment modeling according to the result of the comparison and analysis; The step of obtaining the preset boundary image comprises: The image of the area where the preset boundary is located is acquired through a smartphone; Marking the preset boundary with lines on the acquired image to generate the preset boundary image; The step of extracting features based on the plurality of preset boundary images and obtaining the preset boundary feature distribution comprises: Determine the preset boundary according to the line mark in the preset boundary image; Extracting image features from the outsides of two endpoints of the preset boundary in the preset boundary image to obtain a first preset boundary feature; Extracting image features on both inner and outer sides of the preset boundary in the preset boundary image to obtain a second preset boundary feature; Clearing the line mark, and performing image feature extraction along the preset boundary in the preset boundary image to obtain a third preset boundary feature; The first preset boundary feature, the second preset boundary feature and the third preset boundary feature are combined to generate the preset boundary feature distribution.

2. A path planning method for a cleaning robot as claimed in claim 1, characterized in that: The step of extracting features based on the real-time image and obtaining the feature distribution of the real-time image includes: Analyzing the real-time image to obtain ground information and obstacle information in the real-time image; Performing obstacle recognition and obstacle distance analysis on the obstacle information in the real-time image to generate a first real-time image feature; Performing lateral and longitudinal change analysis on the ground information in the real-time image to generate a second real-time image feature; Performing line segment area analysis on the ground information in the real-time image to generate a third real-time image feature; The ground information, the obstacle information, the first real-time image feature, the second real-time image feature, and the third real-time image feature are combined to generate a real-time image feature distribution.

3. A path planning method for a cleaning robot as claimed in claim 2, characterized in that: The steps of constructing basic environment modeling and the area to be cleaned based on the real-time image feature distribution and the positioning information, and performing path planning according to the area to be cleaned include: Generating relative position relationship information between the feature distributions of the real-time image according to the acquisition time of the real-time image and the acquisition time of the positioning information; Combining the ground information and the obstacle information in the real-time image feature distribution with each other according to the relative position relationship information to generate the basic environment modeling; Analyzing the basic environment modeling to generate an operable area of ​​the cleaning robot; Dividing the operable area into a cleaned area and an area to be cleaned according to the positioning information; Path planning is performed according to the positional relationship between the areas to be cleaned.

4. A path planning method for a cleaning robot as claimed in claim 3, characterized in that: The step of comparing and analyzing the real-time image feature distribution with the preset boundary feature distribution, and setting a virtual obstacle based on the basic environment modeling according to the result of the comparison and analysis includes: Comparing and analyzing the real-time image feature distribution with the preset boundary feature distribution to determine whether the preset boundary exists in the real-time image; According to the result of the determination, the real-time image with the preset boundary is marked as a virtual obstacle image, and the preset boundary in the virtual obstacle image is converted from the ground information to the obstacle information.

5. A path planning device for a cleaning robot, characterized in that: A path planning method for a cleaning robot for implementing any one of claims 1 to 4, comprising: A preset unit is used to receive a plurality of preset boundary images acquired by a smart phone connected to a signal of the cleaning robot, perform feature extraction based on the plurality of preset boundary images, and acquire a preset boundary feature distribution; A driving unit, used to drive a motion device provided on the cleaning robot to enable the cleaning robot to move; A collection unit, used for receiving a real-time image acquired by a camera device arranged on the cleaning robot, performing feature extraction based on the real-time image, and acquiring a feature distribution of the real-time image; A positioning unit, used to receive positioning information acquired by a positioning device provided on the cleaning robot; A planning unit, which constructs basic environment modeling and an area to be cleaned based on the real-time image feature distribution and the positioning information, and performs path planning according to the area to be cleaned; The analysis unit is used to compare and analyze the real-time image feature distribution with the preset boundary feature distribution, and to set virtual obstacles based on the basic environment modeling according to the result of the comparison and analysis.

Images