Map quality detection method, control device, mobile robot and medium

Abstract

The invention discloses a map quality detection method, a control device, a mobile robot and a medium. The method comprises the following steps: acquiring a current map constructed by the robot; pixel matching is carried out on the current map and a historical map to obtain similarity, and / or feature point matching is carried out on the current map and the historical map to obtain feature point pairs, and the historical map is a map constructed according to historical environment data collected by a robot; determining a map folding condition of the current map according to the similarity and / or the feature point pairs; and determining the map quality of the current map according to the map folding condition. In the embodiment of the invention, the quality of the map can be improved, and the situation of map overlapping is avoided.

Description

Technical Field

[0001] This application relates to the field of computers, and in particular to a map quality inspection method, control device, mobile robot, and medium. Background Technology

[0002] In modern robotics, especially in the navigation and localization of mobile robots, it has become standard practice to use LiDAR in conjunction with Simultaneous Localization and Mapping (SLAM) technology to generate high-precision environmental maps. This technology can capture detailed environmental information in real time and create accurate maps based on them, which robots rely on for self-localization and path planning.

[0003] However, SLAM technology faces several challenges in practical applications. One of the most common problems is the "map overlay" phenomenon, which occurs when newly generated map layers are incorrectly superimposed on existing map layers in dynamic environments or due to factors such as sensor errors and software defects, resulting in duplicate or erroneous map information. This overlay phenomenon severely affects map accuracy, thereby impacting the robot's localization accuracy and navigation capabilities. Summary of the Invention

[0004] This application provides a map quality detection method, control device, mobile robot, and medium, which can improve map quality and avoid map overlay.

[0005] In a first aspect, embodiments of this application provide a map quality detection method, including: Get the current map built by the robot; The current map is matched pixel by pixel with the historical map to obtain a similarity, and / or the current map is matched feature point by feature point with the historical map to obtain feature point pairs, wherein the historical map is a map constructed based on the historical environmental data collected by the robot; The overlay status of the current map is determined based on the similarity and / or the feature point pairs; The map quality of the current map is determined based on the overlay situation.

[0006] The map quality detection method according to the first aspect of this application has at least the following beneficial effects: First, the current map constructed by the robot is acquired. In this application embodiment, an environmental map is constructed in real time using data collected by sensors, thereby determining the robot's position on the map. Then, the current map is constructed by pixel matching with the historical environmental data collected by the robot. By comparing pixels, the changes of the current map relative to the historical map can be identified, and the similarity can be obtained. In addition, this application embodiment can also perform feature point matching on the current map and the historical map to obtain feature point pairs, thereby identifying the same features of the current map and the historical map at different scales, and accurately determining the relative position and direction between the maps. After that, the overlay situation of the current map is determined based on the similarity and / or feature point pairs, thereby determining whether the current map has an incorrect overlay situation. Finally, the map quality of the current map is determined based on the overlay situation, achieving accurate judgment of the quality of the current map. This application embodiment achieves pixel-level matching of maps by performing pixel matching between the current map and historical maps, and achieves spatial-level matching of maps by performing feature point matching between the current map and historical maps. Finally, it determines the overlay situation of the current map through similarity and / or feature point pairs, and further detects the map quality of the current map, thereby achieving accurate detection of map quality, avoiding overlay situation, and further improving the robot's positioning accuracy and navigation capability.

[0007] In some embodiments of this application, the step of performing pixel matching between the current map and the historical map to obtain a similarity score includes: The current map is compared with the historical map pixel by pixel to determine the newly added areas and pixel information of the current map relative to the historical map; Based on the newly added areas, non-new areas are determined, and based on the non-new areas, areas of interest are determined; Within the region of interest, the similarity between the current map and the historical map is determined based on the pixel information.

[0008] In the above technical solution, by determining the similarity between the current map and the historical map, the degree of similarity between the current map and the historical map can be determined, thereby enabling rapid detection of map changes and differences, and further enabling the determination of whether the map has an overlay phenomenon, thus improving the accuracy of map quality detection.

[0009] In some embodiments of this application, the step of comparing the current map with the historical map pixel by pixel to determine the newly added areas and pixel information of the current map relative to the historical map includes: Perform pixel traversal on the current map and the historical map to determine the newly added area of ​​the current map relative to the historical map, and determine the first obstacle pixel and the first open area pixel that are the same in the current map and the historical map, as well as the second obstacle pixel and the second open area pixel that are different. The number of pixels in the first obstacle, the number of pixels in the first open area, the number of pixels in the second obstacle, and the number of pixels in the second open area are counted to obtain pixel information.

[0010] In some embodiments of this application, the step of performing feature point matching between the current map and the historical map to obtain feature point pairs includes: Based on a preset visual algorithm, feature points are extracted from the current map and the historical map to obtain the first key feature point of the current map and the second key feature point of the historical map. A brute-force matching process is performed on the first key feature point and the second key feature point to obtain feature point pairs.

[0011] In the above technical solution, by calculating the distance between feature vectors, the closest matching point is found to obtain feature point pairs, thereby enabling the rapid identification of differences and changes between the current map and historical maps.

[0012] In some embodiments of this application, determining the overlay situation of the current map based on the similarity and the feature point pairs includes: A transition matrix is ​​determined based on the feature point pairs, and the transition matrix is ​​used to characterize the offset between the current map and the historical map; The parameters of the transition matrix are extracted to obtain translation parameters and rotation parameters. The translation parameters are used to characterize the translation distance of the current map relative to the historical map, and the rotation parameters are used to characterize the rotation angle of the current map relative to the historical map. The overlay status of the current map is determined based on the similarity, the translation parameter, and the rotation parameter.

[0013] The above technical solution can determine the offset between the current map and the historical map, which facilitates the subsequent correction of the current map and the historical map through the transfer matrix.

[0014] In some embodiments of this application, determining the transition matrix based on the feature point pairs includes: Determine the target feature point pair from the feature point pair, and determine the rotation matrix and translation vector based on the target feature point pair; The objective function is determined based on the rotation matrix, the translation vector, and the target feature point pair; The rotation matrix and translation vector in the objective function are iteratively updated using a preset gradient descent algorithm to obtain the transition matrix.

[0015] In the above technical solution, the accuracy and reliability of the transition matrix can be guaranteed by the transition matrix. Furthermore, the geometric transformation relationship between the current map and the historical map can be described by the transition matrix, which facilitates the subsequent overlapping judgment of the current map.

[0016] In some embodiments of this application, after determining the objective function based on the rotation matrix, the translation vector, and the target feature point pair, the method further includes: The target feature point pairs are verified using a preset iterative algorithm.

[0017] In the above technical solution, the iterative algorithm continuously helps to select and verify target feature point pairs and eliminate outliers, thereby ensuring the robustness of the optimization process and further ensuring the accuracy and reliability of the final transition matrix.

[0018] In some embodiments of this application, determining the overlay status of the current map based on the similarity, the translation parameter, and the rotation parameter includes: When the similarity is higher than a preset threshold, and the translation parameter and the rotation parameter are within the preset range, it is determined that the current map does not have an overlay phenomenon; or, When the similarity is lower than a preset threshold, and / or one of the translation parameter and the rotation parameter exceeds a preset range, it is determined that the current map has an overlay phenomenon.

[0019] In the above technical solution, by comparing the similarity with a preset threshold, and comparing the translation and rotation parameters with a preset range, the overlay situation of the current map can be determined, and it can be further determined whether there is an incorrect overlay phenomenon in the current map.

[0020] In some embodiments of this application, determining the map quality of the current map based on the overlay situation includes: If the current map does not exhibit any overlay phenomenon, it is determined that the current map is a high-quality map; or, If the current map exhibits overlay, it is determined to be a low-quality map.

[0021] The above technical solution enables map quality detection, allowing for timely assessment of map quality.

[0022] In some embodiments of this application, after determining that the current map is a low-quality map, the method further includes: Send a prompt message to the user and receive a decision signal generated by the user based on the prompt message; In response to the decision signal, the historical map is used as the current map or the current map is reconstructed.

[0023] The above technical solution can restore high-quality maps in a timely manner, avoid positioning and navigation on chaotic maps, and avoid long-term failures and downtime.

[0024] In some embodiments of this application, the step of performing brute-force matching on the first key feature point and the second key feature point to obtain feature point pairs includes: Generate a first descriptor corresponding to the first key feature point, and generate a second descriptor corresponding to the second key feature point; Similarity calculation is performed on the first descriptor and the second descriptor to obtain multiple pairs of feature descriptor groups; For each pair of feature descriptor groups, determine the distance value between the first descriptor and the second descriptor; The distance values ​​are sorted to obtain a distance sequence; Select a target feature description subgroup with a preset ranking from the distance sequence, and use the first key feature point and the second key feature point corresponding to the target feature description subgroup as a feature point pair.

[0025] In the above technical solution, the selection of feature point pairs can be achieved by brute-force matching, which can find the closest matching point simply and intuitively without the need for a complex index structure.

[0026] In some embodiments of this application, obtaining the current map constructed by the robot includes: Acquire environmental data collected by the robot; The current map is constructed based on the environmental data.

[0027] In the above technical solution, the robot's current environment, current position, obstacle information in the current environment, etc. can be determined through environmental data, and then a current map can be constructed based on the environmental data, thereby improving the accuracy and robustness of the map.

[0028] Secondly, embodiments of this application also provide a control device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the map quality detection method as described in the first aspect.

[0029] Thirdly, embodiments of this application also provide a mobile robot, including a camera and / or lidar, and a control device as described in the second aspect.

[0030] Fourthly, embodiments of this application also provide a computer-readable storage medium storing computer-executable instructions for causing a computer to perform the map quality detection method as described in the first aspect.

[0031] Other features and advantages of this application will be set forth in the following description and will be apparent in part from the description or may be learned by practicing the application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the description and the accompanying drawings. Attached Figure Description

[0032] The accompanying drawings are used to provide a further understanding of the technical solutions of this application and constitute a part of the specification. They are used together with the embodiments of this application to explain the technical solutions of this application and do not constitute a limitation on the technical solutions of this application.

[0033] Figure 1 This is a flowchart of a map quality detection method provided in one embodiment of this application; Figure 2 This is a flowchart of pixel matching between the current map and historical maps provided in one embodiment of this application; Figure 3 This is a flowchart of an embodiment of the present application for determining the newly added area and pixel information of the current map relative to the historical map; Figure 4 This is a flowchart of feature point matching between the current map and historical maps provided in one embodiment of this application; Figure 5 This is a flowchart of an embodiment of the present application providing a method for determining the overlay of the current map based on similarity and feature point pairs; Figure 6 This is a flowchart of determining the transition matrix based on feature point pairs provided in one embodiment of this application; Figure 7 This is a flowchart of a map quality detection method provided in another embodiment of this application; Figure 8 This is a flowchart of an embodiment of the present application providing a process for determining the overlay status of the current map based on similarity, translation parameters, and rotation parameters; Figure 9 This is a flowchart of determining the map quality of the current map based on the overlay situation, provided in one embodiment of this application; Figure 10 This is a flowchart of a map quality detection method provided in another embodiment of this application; Figure 11 This is a flowchart of a brute-force matching process for a first key feature point and a second key feature point, provided in one embodiment of this application. Figure 12 This is a schematic diagram of a control device provided in one embodiment of this application. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. Furthermore, the features, operations, or characteristics described in the specification can be combined in any suitable manner to form various implementations. Simultaneously, the steps or actions described in the method description can be rearranged or adjusted in a manner readily apparent to those skilled in the art. Therefore, the various orders in the specification and drawings are merely for the clear description of a particular embodiment and do not imply a mandatory order, unless otherwise stated that a particular order must be followed.

[0035] In the description of this application, "several" means one or more, "more than" means two or more, "greater than," "less than," and "exceeding" are understood to exclude the stated number, while "above," "below," and "within" are understood to include the stated number. The use of "first" and "second" in the description is merely for distinguishing technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the order of the indicated technical features.

[0036] The serial numbers assigned to components in this document, such as "first" and "second," are used only to distinguish the described objects and have no sequential or technical meaning. The terms "connection" and "linkage" used in this application, unless otherwise specified, include both direct and indirect connections (linkages).

[0037] In the description of this application, unless otherwise expressly defined, terms such as "setup," "installation," and "connection" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this application in conjunction with the specific content of the technical solution.

[0038] In modern robotics, especially in the navigation and localization of mobile robots, it has become standard practice to use LiDAR in conjunction with Simultaneous Localization and Mapping (SLAM) technology to generate high-precision environmental maps. This technology can capture detailed environmental information in real time and create accurate maps based on them, which robots rely on for self-localization and path planning.

[0039] However, SLAM technology faces several challenges in practical applications. One of the most common problems is the "map overlay" phenomenon, which occurs when newly generated map layers are incorrectly superimposed on existing map layers in dynamic environments or due to factors such as sensor errors and software defects, resulting in duplicate or erroneous map information. This overlay phenomenon severely affects map accuracy, thereby impacting the robot's localization accuracy and navigation capabilities.

[0040] While existing SLAM systems typically include some basic error detection and correction mechanisms, these systems often rely on manual intervention or only trigger processing flows after navigation errors occur. Furthermore, existing error detection methods largely depend on real-time data rather than historical data analysis, which limits the system's ability to prevent and promptly correct overlay errors.

[0041] Based on the above, this application provides a map quality detection method, control device, mobile robot, and medium. First, the current map constructed by the robot is acquired. This application constructs an environmental map in real time using data collected by sensors, thereby determining the robot's position on the map. Then, the current map is matched pixel by pixel with historical environmental data collected by the robot. By comparing pixels, changes in the current map relative to the historical map can be identified, and similarity can be obtained. In addition, this application can also perform feature point matching between the current map and the historical map to obtain feature point pairs, thereby identifying the same features of the current map and the historical map at different scales, and accurately determining the relative position and direction between the maps. Then, the overlay situation of the current map is determined based on the similarity and / or feature point pairs, thereby determining whether the current map has an incorrect overlay situation. Finally, the map quality of the current map is determined based on the overlay situation, achieving accurate judgment of the current map quality. This application embodiment achieves pixel-level matching of maps by performing pixel matching between the current map and historical maps, and achieves spatial-level matching of maps by performing feature point matching between the current map and historical maps. Finally, it determines the overlay situation of the current map through similarity and / or feature point pairs, and further detects the map quality of the current map, thereby achieving accurate detection of map quality, avoiding overlay situation, and further improving the robot's positioning accuracy and navigation capability.

[0042] The embodiments of this application will be further described below with reference to the accompanying drawings.

[0043] like Figure 1 As shown, Figure 1 This is a flowchart of a map quality detection method provided in one embodiment of this application. The map quality detection method in this embodiment includes, but is not limited to, steps S101 to S104.

[0044] Step S101: Obtain the current map constructed by the robot.

[0045] It should be noted that the robot in this embodiment may be equipped with sensors such as lidar, wheel speed odometer, and IMU (Inertial Measurement Unit) to collect environmental data.

[0046] In step S101 of some embodiments, environmental data collected by the robot through sensors is acquired, thereby enabling the robot to determine its current environment, current position, obstacle information in the current environment, etc., and then construct a current map based on the environmental data to improve the accuracy and robustness of the map.

[0047] It is worth noting that in the process of constructing the current map based on environmental data, the embodiments of this application can first preprocess the environmental data, including filtering, denoising, and feature extraction, to improve data quality. Then, feature points are extracted from the preprocessed data, such as feature points in the LiDAR point cloud and visual features in camera images. Next, the sensor data is correlated, that is, data acquired at different times and locations are matched and calibrated. For example, continuous LiDAR scans or camera image sequences are time-aligned, or data from different sensors is fused to improve data consistency and accuracy. Finally, the sensor data is used to estimate the robot's motion state, including position and orientation, and then the current map is constructed based on the sensor data and motion estimation.

[0048] Step S102: Perform pixel matching between the current map and the historical map to obtain similarity, and / or perform feature point matching between the current map and the historical map to obtain feature point pairs.

[0049] It should be noted that the historical map is a map constructed based on historical environmental data collected by the robot.

[0050] In step S102 of some embodiments, the current map is pixel-matched with the historical map constructed from historical environmental data collected by the robot. That is, the pixels of the current map and the historical map are traversed one by one to determine the pixel changes of the current map relative to the historical map, obtain the similarity, and determine the degree of similarity between the current map and the historical map. This facilitates the comparison of the differences between the current map and the historical map at the pixel level. In addition, the embodiments of this application can also perform feature point matching on the current map and the historical map, so as to quickly detect changes and differences in the map and obtain feature point pairs, which facilitates the comparison of the differences between the current map and the historical map at the spatial level.

[0051] It is worth noting that the robot in this application embodiment will dynamically update the map during actual work, that is, collect new environmental data. After the map is built, this application embodiment will back up and store the built map as a historical map. Similarly, the previously collected environmental data will be used as historical environmental data, and the updated map will be used as the current map.

[0052] It is understood that the embodiments of this application can collect environmental data at preset intervals, that is, update the historical map at preset intervals, for example, collect data every two hours, collect data every hour, etc. The embodiments of this application do not impose specific limitations.

[0053] Step S103: Determine the overlay status of the current map based on similarity and / or feature point pairs.

[0054] In step S103 of some embodiments, the overlay situation of the current map is determined based on similarity and / or feature point pairs. That is, the embodiments of this application may determine the overlay situation of the current map based only on similarity or feature points, or may determine the overlay situation of the current map by combining similarity and feature points, thereby determining whether the current map has an incorrect overlay situation.

[0055] Step S104: Determine the map quality of the current map based on the overlay situation.

[0056] In step S104 of some embodiments, the map quality of the current map is determined based on the overlay situation, so as to achieve accurate judgment of map quality and be able to determine whether the map is a high-quality map or a low-quality map.

[0057] Reference Figure 2 , Figure 2 This is a flowchart illustrating pixel-by-pixel matching between the current map and historical maps, provided in one embodiment of this application. The method includes, but is not limited to, steps S201 to S203.

[0058] Step S201: Compare the current map with the historical map pixel by pixel to determine the newly added areas and pixel information of the current map relative to the historical map.

[0059] Step S202: Determine the non-new regions based on the newly added regions, and determine the regions of interest based on the non-new regions.

[0060] Step S203: Within the region of interest, determine the similarity between the current map and the historical map based on pixel information.

[0061] In steps S201 to S203 of some embodiments, during the pixel matching process of the current map and the historical map, firstly, the current map is compared pixel by pixel with the historical map constructed from historical environmental data collected by the robot. This involves traversing each pixel of the current map and the historical map to determine the newly added areas and pixel information of the current map relative to the historical map. This allows for the determination of pixel changes in the current map relative to the historical map. The pixel information can be used to identify the same and different pixels in the current map and the historical map. Then, based on the newly added areas, the non-newly added areas of the current map relative to the historical map are determined. The Region of Interest (ROI) is then determined based on the non-newly added areas. By focusing only on the non-newly added areas, the workload and complexity of the comparison can be greatly reduced. This facilitates subsequent analysis of the degree of change of the current map relative to the historical map in the ROI by combining information such as the newly added area and the total map area. Within the ROI, the similarity between the current map and the historical map is determined based on the pixel information. This determination of the similarity between the current map and the historical map enables rapid detection of map changes and differences. Furthermore, it allows for the determination of whether map overlay occurs, improving the accuracy of map quality detection.

[0062] It is understood that the embodiments of this application only focus on non-new regions and ignore new regions, thereby reducing the workload and complexity of comparison and speeding up the comparison process.

[0063] In some embodiments, in the process of determining the similarity between the current map and the historical map based on pixel information, this application embodiment determines the similarity between the current map and the historical map by calculating the intersection / union ratio or the structural similarity index (SSIM). The process of determining the similarity between the current map and the historical map will be described in detail below, and will not be repeated here.

[0064] Reference Figure 3 , Figure 3 This is a flowchart illustrating the process of determining newly added areas and pixel information in the current map relative to the historical map, as provided in one embodiment of this application. The method includes, but is not limited to, steps S301 to S302.

[0065] Step S301: Perform pixel traversal on the current map and the historical map to determine the newly added area in the current map relative to the historical map, and determine the first obstacle pixel and the first open area pixel that are the same in the current map and the historical map, as well as the second obstacle pixel and the second open area pixel that are different.

[0066] Step S302: Count the number of pixels of the first obstacle, the number of pixels of the first open area, the number of pixels of the second obstacle, and the number of pixels of the second open area to obtain pixel information.

[0067] It is worth noting that the current map and historical map in this application embodiment are both ternary maps, that is, the images include unknown areas, obstacle entities and open areas.

[0068] In steps S301 to S302 of some embodiments, during the pixel comparison between the current map and the historical map, the pixels of the current map and the historical map are first traversed, that is, each pixel of the current map and the historical map is compared. The pixel values ​​at corresponding positions in the current map and the historical map are compared, thereby determining the location areas of the current map and the historical map, the obstacle entities and open areas of the two maps. Furthermore, the first obstacle pixels and the first open area pixels that are the same in the current map and the historical map, and the second obstacle pixels and the second open area pixels that are different, are determined, and the newly added area of ​​the current map relative to the historical map is determined. Then, the number of pixels of each category is counted, that is, the number of first obstacle pixels, the number of first open area pixels, the number of second obstacle pixels and the number of second open area pixels are counted to obtain pixel information. This allows for the differentiation of different types of differences, accurate location and recording of the position of the newly added area, and helps to understand the specific differences between the two maps.

[0069] It is worth noting that, in this embodiment of the application, the same pixel value of an obstacle entity is determined as the first obstacle pixel, and the different pixel value of an obstacle entity is determined as the second obstacle pixel. The same pixel value of an open area is determined as the first open area pixel, and the different pixel value of an open area is determined as the second open area pixel. If no area or pixel matching the current map is found in the historical map, the area is determined as a newly added area of ​​the current map relative to the historical map.

[0070] It should be noted that, in the process of traversing the current map and historical map pixels in this embodiment, for each pixel, its 8 neighboring pixels are considered, rather than the correspondence of absolute positions. The differences are accurately detected by comparing pixels one by one and the 8 neighboring positions.

[0071] In some embodiments of this application, the similarity between the current map and historical maps can be determined by the following formula: ; in, The number of pixels of the first obstacle. This represents the number of pixels in the first open area. The number of pixels of the second obstacle. This represents the number of pixels in the second open area. It is a weight parameter, and its value is between 0 and 1.

[0072] Reference Figure 4 , Figure 4 This is a flowchart illustrating feature point matching between the current map and historical maps, provided in one embodiment of this application. The method includes, but is not limited to, steps S401 to S402.

[0073] Step S401: Based on a preset visual algorithm, feature points are extracted from the current map and the historical map to obtain the first key feature point of the current map and the second key feature point of the historical map.

[0074] Step S402: Perform brute-force matching on the first and second key feature points to obtain feature point pairs.

[0075] In steps S401 to S402 of some embodiments, during the feature point matching process between the current map and the historical map, the embodiments of this application first extract feature points from the current map and the historical map based on a preset visual algorithm to obtain the first key feature point of the current map and the second key feature point of the historical map. The visual algorithm can quickly extract feature points, and then brute-force matching is performed on the first key feature point and the second key feature point. By calculating the distance between the feature vectors, the closest matching point is found to obtain the feature point pair, thereby enabling the rapid identification of the differences and changes between the current map and the historical map.

[0076] It should be noted that the visual algorithm can be the ORB (Oriented Fast and Rotated BRIEF) algorithm. The ORB algorithm has rotation invariance, so even if the viewpoint or pose in the map changes, it can still extract similar feature points. Furthermore, the visual algorithm in this embodiment can also be the SIFT (Scale-Invariant Feature Transform) algorithm, the SURF (Speeded Up Robust Features) algorithm, etc., and this embodiment does not impose specific limitations.

[0077] Reference Figure 5 , Figure 5 This is a flowchart illustrating the process of determining the overlay of the current map based on similarity and feature point pairs, according to one embodiment of this application. The method includes, but is not limited to, steps S501 to S503.

[0078] Step S501: Determine the transition matrix based on the feature point pairs.

[0079] It should be noted that the transition matrix is ​​used to represent the offset between the current map and the historical map.

[0080] In step S501 of some embodiments, a transition matrix is ​​determined based on feature point pairs to ensure the accuracy and reliability of the transition matrix. Furthermore, the transition matrix can be used to describe the geometric transformation relationship between the current map and the historical map, which facilitates subsequent overlap judgment of the current map.

[0081] Step S502: Extract parameters from the transition matrix to obtain translation and rotation parameters.

[0082] It should be noted that the translation parameter is used to characterize the translation distance of the current map relative to the historical map, and the rotation parameter is used to characterize the rotation angle of the current map relative to the historical map.

[0083] In step S502 of some embodiments, parameters are extracted from the transfer matrix to extract the translation and rotation parameters of the transfer matrix. The translation distance and rotation angle between the current map and the historical map can be determined by the translation distance and rotation angle, and the alignment degree of the two maps can be accurately determined.

[0084] Step S503: Determine the overlay status of the current map based on similarity, translation parameters, and rotation parameters.

[0085] In step S503 of some embodiments, the map quality of the current map is determined based on similarity, translation parameters, and rotation parameters. By comparing parameters at the pixel level and spatial level, the overlay situation of the current map is accurately judged, avoiding situations where the map changes too much or there are incorrect overlapping areas, thereby improving the accuracy of the robot navigation path.

[0086] It is worth noting that in the process of determining the overlay status of the current map based on similarity and feature point pairs, the embodiments of this application first determine the transfer matrix based on the feature point pairs, thereby determining the offset between the current map and the historical map. This facilitates the subsequent correction of the current map and the historical map through the transfer matrix. The parameters of the transfer matrix are extracted to obtain translation parameters and rotation parameters, which are used to determine the translation distance and rotation angle of the current map relative to the historical map, thereby describing the spatial relationship between the current map and the historical map. Finally, the overlay status of the current map is determined based on similarity, translation parameters, and rotation parameters.

[0087] In some embodiments, the overlay status of the current map can also be determined based on similarity or feature point pairs. Specifically, when determining the overlay status of the current map based on feature point pairs, the embodiments of this application determine the transition matrix based on the feature point pairs, then extract parameters from the transition matrix to obtain translation parameters and rotation parameters, and finally determine the overlay status of the current map based on similarity, translation parameters, and rotation parameters. When determining the overlay status of the current map based on similarity, the similarity can be directly compared with a preset threshold to determine the overlay status of the current map.

[0088] Reference Figure 6 , Figure 6 This is a flowchart illustrating the determination of a transition matrix based on feature point pairs according to an embodiment of this application. The method includes, but is not limited to, steps S601 to S603.

[0089] Step S601: Determine the target feature point pair from the feature point pair, and determine the rotation matrix and translation vector based on the target feature point pair.

[0090] Step S602: Determine the objective function based on the rotation matrix, translation vector, and target feature point pairs.

[0091] Step S603: Iteratively update the rotation matrix and translation vector in the objective function using a preset gradient descent algorithm to obtain the transition matrix.

[0092] In steps S601 to S603 of some embodiments, during the process of determining the transition matrix based on feature point pairs, target feature point pairs are first selected from the feature point pairs, and rotation matrices and translation vectors are determined based on the target feature point pairs. The rotation matrix is ​​used to characterize the rotation transformation between the target feature point pairs, and the translation vector is used to characterize the translation transformation between the target feature point pairs. Then, an objective function is determined based on the rotation matrix, translation vector, and target feature point pairs. The objective function at this time aims to minimize the sum of squared Euclidean distances between feature point pairs. The rotation matrix and translation vector in the objective function are then iteratively updated using a preset gradient descent algorithm to obtain the transition matrix, thereby ensuring the accuracy and reliability of the transition matrix. Furthermore, the transition matrix can be used to describe the geometric transformation relationship between the current map and the historical map, which is convenient for subsequent overlap judgment of the current map.

[0093] Specifically, the objective function in this application embodiment can be expressed as follows: ; in, Represents the rotation matrix. Represents the translation vector. and These are the matched feature point pairs.

[0094] This embodiment of the application also calculates the gradient of the objective function with respect to the rotation matrix and translation vector, deriving the update formula. For each iteration, the update formula is expressed as follows: ; in, Represents the rotation matrix. Represents the translation vector. This is the learning rate.

[0095] Iterative updates are performed using the gradient descent algorithm. and The process continues until convergence, and each update gradually optimizes the transition matrix parameters by reducing the objective function value.

[0096] Reference Figure 7 , Figure 7 This is a flowchart of a map quality detection method provided in another embodiment of this application. The map quality detection method of this application embodiment includes, but is not limited to, step S701.

[0097] It should be noted that step S701 occurs after the objective function is determined based on the rotation matrix, translation vector, and target feature point pair.

[0098] Step S701: Verify the target feature point pair using a preset iterative algorithm.

[0099] In step S701 of some embodiments, during the gradient descent optimization process, the present application embodiments will also verify the target feature point pairs through a preset iterative algorithm. The iterative algorithm continuously helps to select and verify the target feature point pairs and eliminate outliers, thereby ensuring the robustness of the optimization process and further ensuring the accuracy and reliability of the final transition matrix.

[0100] It should be noted that the iterative algorithm in this embodiment can be Random Sample Consensus (RANSAC), gradient descent, conjugate gradient, etc. In this embodiment, RANSAC is used. This embodiment does not impose specific restrictions on the selection of the iterative algorithm.

[0101] Reference Figure 8 , Figure 8 This is a flowchart illustrating the process of determining the overlay of the current map based on similarity, translation parameters, and rotation parameters, according to one embodiment of this application. The method includes, but is not limited to, steps S801 to S802.

[0102] Step S801: When the similarity is higher than the preset threshold and the translation and rotation parameters are within the preset range, it is determined that there is no overlay phenomenon in the current map.

[0103] In step S801 of some embodiments, during the process of determining the map quality of the current map, the embodiments of this application will set a preset range in advance and compare the similarity with a preset threshold, and compare the translation parameters and rotation parameters with the preset range to determine the overlay situation of the current map. It can further determine whether there is an incorrect overlay phenomenon in the current map. Specifically, when the similarity is higher than the preset threshold and the translation parameters and rotation parameters do not exceed the preset range, it indicates that the similarity between the current map and the historical map is high and the overlap between the current map and the historical map is high. It can be determined that there is no overlay phenomenon in the current map, thus realizing the detection of map overlay.

[0104] It should be noted that the embodiments of this application can accurately determine the alignment degree of two maps by quantifying the translation distance and rotation angle. When comparing historical maps and current maps, changes in translation distance and rotation angle can indicate changes in the environment. If the translation distance is small but the rotation angle changes, it may mean that the environment has changed direction or rotated. If both the translation distance and rotation angle are small, the maps can be considered to be highly overlapping; otherwise, there may be significant differences.

[0105] Step S802: When the similarity is lower than a preset threshold, and / or one of the translation parameter and rotation parameter exceeds a preset range, it is determined that there is an overlay phenomenon in the current map.

[0106] In step S802 of some embodiments, if the similarity is lower than a preset threshold, it indicates that the similarity between the current map and the historical map is low and the difference between the two maps is large. If either the translation parameter or the rotation parameter exceeds a preset range, it indicates that there is a certain transformation between the current map and the historical map, and incorrect overlap may occur. If any of the above three situations occur, it is determined that the current map has an overlay phenomenon, thereby achieving accurate detection of map overlay, avoiding excessive map changes or overlapping areas, and improving the accuracy of robot navigation paths.

[0107] It is worth noting that the overlay phenomenon in this application embodiment refers to the phenomenon where the newly generated map layer is incorrectly superimposed on the original map layer, resulting in duplicate or incorrect coverage of map information. If any of the following occurs: similarity is lower than a preset threshold, translation parameters exceed a preset range, or rotation parameters exceed a preset range, this application embodiment will determine that the current map has an overlay phenomenon.

[0108] It should be noted that the preset threshold in this application embodiment can be set according to the user's needs or the robot's working parameters. Similarly, the preset range in this application embodiment can be set according to the user's needs or the robot's working parameters. This application embodiment does not impose any specific limitations.

[0109] Reference Figure 9 , Figure 9 This is a flowchart illustrating the process of determining the map quality of the current map based on overlay conditions, provided in one embodiment of this application. The method includes, but is not limited to, steps S803 to S804.

[0110] Step S803: If there is no overlay phenomenon in the current map, the current map is determined to be a high-quality map.

[0111] Step S804: If the current map has an overlay phenomenon, determine that the current map is a low-quality map.

[0112] In steps S803 to S804 of some embodiments, if there is no overlay phenomenon in the current map, it means that the layers of the current map and the historical map can be correctly overlaid and the information of the two maps is accurately covered, and the current map can be directly determined to be a high-quality map; if there is an overlay phenomenon in the current map, it means that the layers of the current map and the historical map cannot be correctly overlaid, and there may be information duplication or incorrect coverage, and the current map can be directly determined to be a low-quality map.

[0113] Reference Figure 10 , Figure 10 This is a flowchart of a map quality detection method provided in another embodiment of this application. The map quality detection method of this application includes, but is not limited to, steps S901 to S902.

[0114] It should be noted that steps S901 to S902 occur after it is determined that the current map is a low-quality map.

[0115] Step S901: Send a prompt message to the user to receive the decision signal generated by the user based on the prompt message.

[0116] Step S902: In response to the decision signal, use the historical map as the current map or reconstruct the current map.

[0117] In steps S901 to S902 of some embodiments, after determining that the current map is a low-quality map, this application embodiment will send a prompt message to the user, such as a voice prompt message, an indicator light prompt message, etc., so that the user can respond after receiving the prompt message, thereby enabling the user to understand the status of the map quality in a timely manner and take corresponding measures. The application embodiment receives the decision signal generated by the user based on the prompt message. After receiving the decision signal generated by the user, this application embodiment responds to the decision signal by using the historical map as the current map or reconstructing the current map, that is, restoring the current map to the historical map or reconstructing a new map, thereby enabling the timely restoration of a high-quality map, avoiding positioning and navigation on a chaotic map, and avoiding long-term failures and downtime.

[0118] It is understood that the embodiments of this application can send prompts to users through operation pages, displays, mini-programs, etc. on the terminal, and users can make decisions manually or automatically based on the prompts. The embodiments of this application do not impose specific limitations.

[0119] Reference Figure 11 , Figure 11 This is a flowchart illustrating a brute-force matching process for a first key feature point and a second key feature point, provided in one embodiment of this application. The method includes, but is not limited to, steps S1001 to S1005.

[0120] Step S1001: Generate a first descriptor corresponding to the first key feature point, and generate a second descriptor corresponding to the second key feature point.

[0121] Step S1002: Calculate the similarity between the first descriptor and the second descriptor to obtain multiple pairs of feature descriptor groups.

[0122] Step S1003: For each pair of feature descriptors, determine the distance value between the first descriptor and the second descriptor.

[0123] Step S1004: Sort the distance values ​​to obtain a distance sequence.

[0124] Step S1005: Select a target feature description subgroup with a preset ranking in the distance sequence, and take the first key feature point and the second key feature point corresponding to the target feature description subgroup as a feature point pair.

[0125] In steps S1001 to S1005 of some embodiments, during the brute-force matching of the first key feature point and the second key feature point, this embodiment first generates a first descriptor corresponding to the first key feature point and a second descriptor corresponding to the second key feature point. The descriptor is a set of numerical values ​​that can describe the local image content around the feature point. Then, similarity calculation is performed on the first and second descriptors to find the best match. The similarity between each pair of feature point descriptors is compared to obtain multiple pairs of feature descriptor groups. For each pair of feature descriptor groups, the distance value between the first and second descriptors is determined. Specifically, this embodiment can calculate the distance value between the first and second descriptors using Euclidean distance, Hamming distance, cosine similarity, etc. The distances of each pair of feature point descriptors are then sorted to obtain a distance sequence. Finally, a target feature descriptor group with a preset ranking is selected from the distance sequence, and the first and second key feature points corresponding to the target feature descriptor group are used as feature point pairs. This achieves the selection of feature point pairs. Brute-force matching can simply and intuitively find the closest matching point without requiring a complex index structure.

[0126] It should be noted that the preset ranking in this application embodiment can be set according to the user's needs or the sorting rules of the distance sequence. For example, if the distance value of each pair of feature descriptors in the distance sequence decreases step by step, then this application embodiment can select the N pairs of feature descriptors ranked at the bottom of the distance sequence as the target feature descriptors. Similarly, if the distance value of each pair of feature descriptors in the distance sequence increases step by step, then this application embodiment can select the N pairs of feature descriptors ranked at the top of the distance sequence as the target feature descriptors. This application embodiment does not impose specific restrictions on the selection of the target feature descriptors.

[0127] To more clearly explain the map quality inspection method, control device, mobile robot, and medium described above, specific examples are provided below.

[0128] Example 1: Example 1 is a specific example illustrating the map quality detection method of this application.

[0129] Step 1: The robot collects environmental data using sensors such as LiDAR, wheel speed odometer, and IMU. Then, it constructs a map using the SLAM algorithm. Specifically, the SLAM algorithm processes the sensor data to build a SLAM map of the environment.

[0130] Understandably, after the map is built, it is backed up and stored as a historical map or a reference map.

[0131] During actual operation, the robot dynamically updates the map based on sensor data, during which the map undergoes various environmental changes. After one work cycle, the updated map is called the current map.

[0132] It should be noted that the work cycle in this example can be the cleaning cycle of a robot vacuum cleaner, the transportation cycle of an AGV robot, etc., and both the current map and the backup map are ternary maps, that is, gray represents unknown areas, black represents obstacle entities, and white represents open areas.

[0133] Step 2: Perform pixel comparison and quality checks on the current map and historical maps.

[0134] Specifically, the current map and the backup map are compared pixel by pixel at corresponding locations to count the number and location of pixels in the newly added area (new_area) in the current map relative to the backup map, and the number of pixels with the same hit (new_area). The same number of Miss pixels ( Different numbers of Hit pixels ( Different numbers of Miss pixels Furthermore, the corresponding pixel position mentioned above refers to the corresponding position within an 8-neighborhood, not the corresponding position at an absolute location.

[0135] Based on the number and location information of pixels (new_area) in the newly added area, determine whether the current map is in the non-new area relative to the backup map. Then, based on the pixel information obtained above ( , , , ) Calculate the similarity between the current map and historical maps.

[0136] Step 3: Perform feature point matching and transition matrix estimation on the current map and historical maps.

[0137] Specifically, key feature points are extracted from the current map and the backup map respectively. Then, brute-force matching is performed on the extracted feature points to obtain pairs of feature points. Finally, brute-force matching is performed on the extracted feature points to obtain pairs of feature points.

[0138] It is worth noting that the embodiments of this application define an objective function with the goal of minimizing the sum of squared Euclidean distances between feature point pairs.

[0139] The objective function can be expressed as follows: ; in, Represents the rotation matrix. Represents the translation vector. and These are the matched feature point pairs.

[0140] This embodiment of the application also calculates the gradient of the objective function with respect to the rotation matrix and translation vector, deriving the update formula. For each iteration, the update formula is expressed as follows: ; in, Represents the rotation matrix. Represents the translation vector. This is the learning rate.

[0141] Iterative updates are performed using the gradient descent algorithm. and The process continues until convergence, and each update gradually optimizes the transition matrix parameters by reducing the objective function value.

[0142] During gradient descent optimization, the RANSAC algorithm continuously helps select and verify feature matching pairs, eliminate outliers, and ensure the robustness of the optimization process. The continuous involvement of RANSAC ensures the accuracy and reliability of the final transition matrix.

[0143] Step 4: Extract the translation distance and rotation angle.

[0144] Specifically, through the optimized transition matrix Extract translation parameters respectively and rotation parameters Translation parameters This represents the translation distance and rotation parameters of the map. Indicates the rotation angle of the map.

[0145] Step 5: Perform an overlay judgment on the current map.

[0146] Specifically, the overlap between two maps is determined based on the translation distance and rotation angle. The map quality of the current map is evaluated by calculating the overlap of the two maps on feature point pairs.

[0147] It is worth noting that this example also has an exception handling and feedback mechanism. If an overlay phenomenon is detected in the current map, the system will notify the user or operator through the APP (Application, mobile software) or other means.

[0148] Operators can make decisions manually or automatically based on the prompts, such as restoring a historical map to the current map or rebuilding a new map.

[0149] By promptly restoring high-quality maps, we can avoid positioning and navigation on chaotic maps, thus preventing unpredictable outcomes and system failures.

[0150] In this example, the quality of the backup map is ensured by backing it up after SLAM map construction is completed and at the end of each work cycle. A comparison check is performed after each map update; if the current map is evaluated as high quality, the backup map is overwritten. This ensures that the currently used map is always of high quality, and especially avoids overlaying. A dual detection mechanism using pixel-level matching and feature point matching can promptly detect abnormal changes in the map, such as overlaying or other quality issues. These are then assessed using set thresholds to ensure the map used by the robot during operation has high accuracy and consistency, thereby improving the precision of localization and navigation.

[0151] Furthermore, when map quality issues are detected, this embodiment can automatically execute preset countermeasures, such as notifying the operator to perform a manual check or automatically restoring to a historical map with better quality. The restoration mechanism ensures that the system can quickly return to normal operation when map problems occur, avoiding prolonged downtime and failures. This automated processing mechanism greatly improves system reliability and efficiency, reducing delays and errors caused by manual intervention. Through real-time feedback and automated processing, users or operators can promptly understand the map quality status and take appropriate measures based on the prompts. This instant feedback mechanism increases user trust and satisfaction with the system, making operation simpler.

[0152] like Figure 12 As shown, Figure 12 This is a schematic diagram of a control device provided in one embodiment of this application.

[0153] The control device 1000 in this embodiment includes one or more processors 1001 and a memory 1002. Figure 12 The example uses a processor 1001 and a memory 1002.

[0154] Processor 1001 and memory 1002 can be connected via a bus or other means. Figure 12 Taking the example of a connection between China and Israel via a bus.

[0155] Memory 1002, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory 1002 may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory 1002 may optionally include memory 1002 remotely located relative to processor 1001, and these remote memories can be connected to the control device 1000 via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0156] Those skilled in the art will understand that Figure 12 The device structure shown does not constitute a limitation on the control device 1000, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0157] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network nodes. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0158] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0159] The non-transitory software program and instructions required to implement the map quality detection method of the above embodiments are stored in memory. When executed by the processor, the map quality detection method of the above embodiments is executed.

[0160] In some embodiments, this application also provides a mobile robot. Since the mobile robot of this application has a camera and / or lidar and the control device of the above embodiments, and the control device of the above embodiments can execute the map quality detection method of the above embodiments, the specific implementation and technical effects of the mobile robot of this application can refer to the specific implementation and technical effects of the map quality detection method of any of the above embodiments.

[0161] The device or system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0162] Furthermore, one embodiment of this application provides a computer-readable storage medium storing computer-executable instructions that are executed by a processor or controller.

[0163] It will be understood by those skilled in the art that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include computer-readable storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer-readable storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer-readable storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, as is known to those skilled in the art, communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0164] The above provides a detailed description of the preferred embodiments of this application. However, this application is not limited to the above-described embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of this application. All such equivalent modifications or substitutions are included within the scope defined by this application.

Claims

1. A map quality detection method characterized by comprising: The method comprises: acquiring a current map constructed by a robot; performing pixel matching on the current map and a historical map to obtain a similarity, and / or performing feature point matching on the current map and the historical map to obtain a feature point pair, wherein the historical map is a map constructed according to historical environment data collected by the robot; determining an overlay situation of the current map according to the similarity and / or the feature point pair; determining a map quality of the current map according to the overlay situation.

2. The map quality detection method according to claim 1, characterized by, The pixel matching comprises: comparing pixels of the current map with pixels of the historical map to determine an added area of the current map relative to the historical map and pixel information; determining a non-added area according to the added area, and determining a region of interest according to the non-added area; determining the similarity between the current map and the historical map in the region of interest according to the pixel information.

3. The map quality detection method according to claim 2, characterized by, The pixel comparison comprises: performing pixel traversal on the current map and the historical map to determine the added area of the current map relative to the historical map, and to determine first obstacle pixels and first empty area pixels that are the same in the current map and the historical map, and second obstacle pixels and second empty area pixels that are different; counting the number of the first obstacle pixels, the number of the first empty area pixels, the number of the second obstacle pixels and the number of the second empty area pixels to obtain the pixel information.

4. The map quality detection method according to claim 1, characterized by, The feature point matching comprises: extracting first key feature points of the current map and second key feature points of the historical map based on a preset visual algorithm; performing brute-force matching on the first key feature points and the second key feature points to obtain the feature point pair.

5. The map quality detection method according to claim 1, characterized by, The determination of the overlay situation comprises: determining a transfer matrix according to the feature point pair, the transfer matrix being used to represent an offset between the current map and the historical map; extracting parameters from the transfer matrix to obtain a translation parameter and a rotation parameter, wherein the translation parameter is used to represent a translation distance of the current map relative to the historical map, and the rotation parameter is used to represent a rotation angle of the current map relative to the historical map; determining the overlay situation of the current map according to the similarity, the translation parameter and the rotation parameter.

6. The map quality detection method according to claim 5, characterized by, The determination of the transfer matrix comprises: determining a target feature point pair in the feature point pair, and determining a rotation matrix and a translation vector according to the target feature point pair; determining a target function according to the rotation matrix, the translation vector and the target feature point pair; The rotation matrix and the translation vector in the target function are iteratively updated by a preset gradient descent algorithm to obtain a transition matrix.

7. The map quality detection method according to claim 6, characterized by, After determining the target function according to the rotation matrix, the translation vector and the target feature point pair, the method further comprises: The target feature point pair is verified by a preset iterative algorithm.

8. The map quality detection method according to claim 5, characterized by, The determination of the overlay condition of the current map according to the similarity, the translation parameter and the rotation parameter comprises: When the similarity is higher than a preset threshold, and the translation parameter and the rotation parameter are both within a preset range, it is determined that the current map does not have an overlay phenomenon; Or, When the similarity is lower than a preset threshold, and / or one of the translation parameter and the rotation parameter is out of a preset range, it is determined that the current map has an overlay phenomenon.

9. The map quality detection method according to claim 8, characterized by, The determination of the map quality of the current map according to the overlay condition comprises: If the current map does not have an overlay phenomenon, the current map is determined to be a high-quality map; Or, If the current map has an overlay phenomenon, the current map is determined to be a low-quality map.

10. The map quality detection method according to claim 9, characterized by, After determining that the current map is a low-quality map, the method further comprises: sending prompt information to a user to receive a decision signal generated by the user according to the prompt information; in response to the decision signal, taking the historical map as a current map or reconstructing a current map.

11. The map quality detection method according to claim 4, characterized by, The brute-force matching of the first key feature point and the second key feature point to obtain a feature point pair comprises: generating a first descriptor corresponding to the first key feature point and a second descriptor corresponding to the second key feature point; performing similarity calculation on the first descriptor and the second descriptor to obtain a plurality of feature descriptor groups; for each feature descriptor group, determining a distance value between the first descriptor and the second descriptor; sorting the distance values to obtain a distance sequence; selecting a target feature descriptor group with a preset ranking in the distance sequence, and taking the first key feature point and the second key feature point corresponding to the target feature descriptor group as a feature point pair.

12. The map quality detection method according to claim 1, characterized by, The acquisition of the current map constructed by the robot comprises: acquiring environment data collected by the robot; constructing a current map according to the environment data.

13. A control device comprising: A memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the map quality detection method according to any one of claims 1 to 12 when executing the computer program.

14. A mobile robot, characterized by A control device comprising a camera and / or a laser radar and the control device according to claim 13.

15. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer executable instructions for causing a computer to execute the map quality detection method according to any one of claims 1 to 12.

Images