Search

CN-114494278-B - Map partitioning and construction, object identification and cleaning method, device and storage medium

CN114494278BCN 114494278 BCN114494278 BCN 114494278BCN-114494278-B

Abstract

The embodiment of the application provides a map partitioning and constructing method, an object identifying and cleaning method, equipment and a storage medium. In the embodiment of the application, the autonomous mobile equipment traverses the working environment by adopting a plurality of track modes with different main travelling directions, so that the whole working environment can be covered as much as possible, thereby shooting partition reference objects in the working environment as much as possible, increasing the probability of shooting front angle images of the partition reference objects, being beneficial to more accurately identifying the partition reference objects and the positions thereof in the working environment and improving the accuracy of map partition based on the partition reference objects.

Inventors

  • MA XIAOYUE
  • CAI RUIYING
  • BAO LIANG

Assignees

  • 科沃斯机器人股份有限公司

Dates

Publication Date
20260505
Application Date
20201112

Claims (15)

  1. 1. A map partitioning method, comprising: The autonomous mobile equipment is controlled to traverse the working environment by adopting at least two track modes, and a plurality of environment images are acquired in the traversing process by adopting the at least two track modes; Marking the position of a partition reference object in an environment map corresponding to a working environment according to the frequency and the position of the partition reference object in the working environment in the plurality of environment images; Partitioning the environment map according to the position of the partitioning reference object in the environment map, wherein the main advancing directions of the at least two track modes are different, the main advancing directions are advancing directions meeting the set condition, and the cleaning paths of the at least two track modes are overlapped in a staggered mode.
  2. 2. The method of claim 1, wherein controlling the autonomous mobile device to traverse the work environment in at least two trajectory modes comprises: After the autonomous mobile equipment traverses the operation environment by adopting the previous track mode, controlling the autonomous mobile equipment to return to the appointed starting position again, and continuously traversing the operation environment again by adopting the next track mode; Or alternatively After the autonomous mobile device traverses the working environment in the last track mode, the autonomous mobile device is controlled to start from the current position, and the next track mode is continuously adopted to traverse the working environment again.
  3. 3. The method of claim 1, wherein marking the location of the partition reference in the environment map corresponding to the work environment based on the number of occurrences and locations of the partition reference in the plurality of environment images, which are present in the work environment, comprises: Selecting a target environment image containing the partition reference object from the plurality of environment images; marking candidate positions of the partition references in the environment map according to the positions of the partition references in the target environment image; and removing candidate positions with the marked times smaller than a set time threshold value to obtain the positions of the partition references in the environment map.
  4. 4. The method of claim 3, wherein selecting a target environmental image containing the partition reference from the plurality of environmental images comprises: And inputting the plurality of environment images into an image recognition model for recognition to obtain a target environment image comprising the partition reference object, wherein the image recognition model is obtained by training in advance with a sample image comprising the partition reference object.
  5. 5. The method of any one of claims 1-4, wherein the at least two trajectory patterns include a horizontal arcuate trajectory pattern and a vertical arcuate trajectory pattern, a main direction of travel of the horizontal arcuate trajectory pattern and the vertical arcuate trajectory pattern being perpendicular to each other, and the main direction of travel being a direction of travel of the autonomous mobile device along a long side trajectory of the arcuate trajectory.
  6. 6. The method according to any one of claims 1 to 4, wherein the at least two track modes comprise a horizontal irregular track mode and a vertical irregular track mode, the travel tracks corresponding to the horizontal irregular track mode and the vertical irregular track mode comprise a straight track and a curve track which are connected, the travel direction of the autonomous mobile device along the straight track is the main travel direction, and the main travel directions of the horizontal irregular track mode and the vertical irregular track mode are mutually perpendicular.
  7. 7. The method according to any one of claims 1-4, wherein the track gap between different track patterns is different and for two adjacent track patterns the track gap of the latter track pattern is larger than the track gap of the former track pattern.
  8. 8. The semantic map construction method is characterized by comprising the following steps of: controlling the autonomous mobile device to traverse the operating environment in at least two track modes; Collecting a plurality of environment images in the traversal process adopting the at least two track modes; And constructing a semantic map of the working environment based on the plurality of environment images, wherein the main advancing directions of at least two track modes are different, the main advancing directions refer to advancing directions meeting set conditions, and cleaning paths of the at least two track modes are overlapped in a staggered mode.
  9. 9. The method of claim 8, wherein constructing a semantic map of a work environment based on the plurality of environmental images comprises: Identifying objects existing in the working environment and semantic information thereof based on the plurality of environment images; determining at least two partitions and semantic information thereof contained in the working environment according to objects and semantic information thereof existing in the working environment; And constructing a semantic map of the working environment according to the at least two partitions and the semantic information thereof.
  10. 10. The method of claim 9, wherein determining at least two partitions and their semantic information contained in the work environment based on objects and their semantic information present in the work environment comprises: Dividing the working environment into at least two partitions according to partition references and semantic information thereof existing in the working environment; and determining semantic information of the at least two partitions according to the non-partition reference objects and the semantic information thereof contained in the at least two partitions.
  11. 11. An object recognition method, comprising: controlling the autonomous mobile device to traverse the operating environment in at least two track modes; Collecting a plurality of environment images in the traversal process adopting the at least two track modes; And identifying target objects existing in the working environment based on the plurality of environment images, wherein the main traveling directions of the at least two track modes are different, the main traveling directions are traveling directions meeting set conditions, and cleaning paths of the at least two track modes are overlapped in a staggered mode.
  12. 12. An autonomous mobile device is characterized by comprising a device body, wherein a memory and a processor are arranged on the device body; The memory is used for storing a computer program; The processor, coupled to the memory, is configured to execute the computer program for: the autonomous mobile equipment is controlled to traverse the working environment in at least two track modes, and a plurality of environment images are acquired in the traversing process; Marking the position of a partition reference object in an environment map corresponding to a working environment according to the frequency and the position of the partition reference object in the working environment in the plurality of environment images; Partitioning the environment map according to the position of the partitioning reference object in the environment map, wherein the main advancing directions of the at least two track modes are different, the main advancing directions are advancing directions meeting the set condition, and the cleaning paths of the at least two track modes are overlapped in a staggered mode.
  13. 13. An autonomous mobile device is characterized by comprising a device body, wherein a memory and a processor are arranged on the device body; The memory is used for storing a computer program; The processor, coupled to the memory, is configured to execute the computer program for: controlling the autonomous mobile device to traverse the operating environment in at least two track modes; Collecting a plurality of environment images in the traversal process adopting the at least two track modes; And constructing a semantic map of the working environment based on the plurality of environment images, wherein the main advancing directions of at least two track modes are different, the main advancing directions refer to advancing directions meeting set conditions, and cleaning paths of the at least two track modes are overlapped in a staggered mode.
  14. 14. An autonomous mobile device is characterized by comprising a device body, wherein a memory and a processor are arranged on the device body; The memory is used for storing a computer program; The processor, coupled to the memory, is configured to execute the computer program for: controlling the autonomous mobile device to traverse the operating environment in at least two track modes; Collecting a plurality of environment images in the traversal process adopting the at least two track modes; And identifying target objects existing in the working environment based on the plurality of environment images, wherein the main traveling directions of the at least two track modes are different, the main traveling directions are traveling directions meeting set conditions, and cleaning paths of the at least two track modes are overlapped in a staggered mode.
  15. 15. A computer readable storage medium storing a computer program, characterized in that the computer program, when executed by a processor, causes the processor to carry out the steps of the method according to any one of claims 1-11.

Description

Map partitioning and construction, object identification and cleaning method, device and storage medium Technical Field The application relates to the technical field of artificial intelligence, in particular to a map partitioning and constructing method, an object identifying and cleaning method, equipment and a storage medium. Background With the development of artificial intelligence technology, the robot functions are becoming more and more intelligent. For example, the sweeping robot releases the hands of the user, and the sweeping robot can automatically complete floor cleaning in a room to release the hands of the user. The sweeping robot has more and more powerful functions, and the track modes supported by the sweeping robot are more and more abundant, for example, the sweeping robot can support partition sweeping. The robot automatically detects environmental information of the working environment through sensors such as a laser radar and a vision sensor when walking in the environment where the robot is located, constructs an environmental map of the working environment based on the environmental information, and intelligently partitions the environmental map based on the environmental information. However, existing partitions have the problem of being not accurate enough. Disclosure of Invention Aspects of the present application provide a map partitioning and construction, object recognition and cleaning method, apparatus, and storage medium for improving accuracy of environmental map partitioning. The embodiment of the application provides a map partitioning method, which comprises the steps of controlling an autonomous mobile device to traverse a working environment by adopting at least two track modes, collecting a plurality of environment images in the traversing process by adopting the at least two track modes, marking the position of a partitioning reference object in an environment map corresponding to the working environment according to the occurrence times and the position of the partitioning reference object in the plurality of environment images in the working environment, and partitioning the environment map according to the position of the partitioning reference object in the environment map, wherein the main advancing directions of the at least two track modes are different, and the main advancing directions are advancing directions meeting set conditions. The embodiment of the application also provides a semantic map construction method, which comprises the steps of controlling the autonomous mobile equipment to traverse the working environment by adopting at least two track modes, collecting a plurality of environment images in the traversing process by adopting the at least two track modes, constructing a semantic map of the working environment based on the plurality of environment images, wherein main traveling directions of the at least two track modes are different, and the main traveling directions are traveling directions meeting set conditions. The embodiment of the application also provides an object identification method, which comprises the steps of controlling the autonomous mobile equipment to traverse the working environment by adopting at least two track modes, collecting a plurality of environment images in the traversing process by adopting the at least two track modes, identifying a target object existing in the working environment based on the plurality of environment images, wherein main traveling directions of the at least two track modes are different, and the main traveling directions are traveling directions meeting set conditions. The embodiment of the application also provides a cleaning method, which comprises the steps of controlling the cleaning robot to execute the cleaning task in the operation area by adopting the first track mode, and controlling the cleaning robot to continuously execute the cleaning task in the operation area again by adopting the second track mode after completing the cleaning task by adopting the first track mode, wherein the main advancing directions of the first track mode and the second track mode are different, and the main advancing directions are advancing directions meeting the set condition. The embodiment of the application also provides an autonomous mobile device, which comprises a device body, wherein a memory and a processor are arranged on the device body, the memory is used for storing a computer program, the processor is coupled with the memory and used for executing the computer program and used for controlling the autonomous mobile device to traverse a working environment in at least two track modes and collect a plurality of environment images in the traversing process in which the at least two track modes are adopted, the position of a partition reference object corresponding to the working environment is marked in an environment map corresponding to the working environment according to the occurrence time