WO2017166594A1 - 一种室内地图构建方法、装置和存储介质 - Google Patents

一种室内地图构建方法、装置和存储介质 Download PDF

Info

Publication number
WO2017166594A1
WO2017166594A1 PCT/CN2016/096104 CN2016096104W WO2017166594A1 WO 2017166594 A1 WO2017166594 A1 WO 2017166594A1 CN 2016096104 W CN2016096104 W CN 2016096104W WO 2017166594 A1 WO2017166594 A1 WO 2017166594A1
Authority
WO
WIPO (PCT)
Prior art keywords
point cloud
cloud data
laser point
indoor
key frame
Prior art date
Application number
PCT/CN2016/096104
Other languages
English (en)
French (fr)
Inventor
种道晨
周煜远
张宇智
刘玉亭
Original Assignee
百度在线网络技术(北京)有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 百度在线网络技术(北京)有限公司 filed Critical 百度在线网络技术(北京)有限公司
Publication of WO2017166594A1 publication Critical patent/WO2017166594A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/38Electronic maps specially adapted for navigation; Updating thereof
    • G01C21/3804Creation or updating of map data
    • G01C21/3807Creation or updating of map data characterised by the type of data
    • G01C21/383Indoor data
    • GPHYSICS
    • G09EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
    • G09BEDUCATIONAL OR DEMONSTRATION APPLIANCES; APPLIANCES FOR TEACHING, OR COMMUNICATING WITH, THE BLIND, DEAF OR MUTE; MODELS; PLANETARIA; GLOBES; MAPS; DIAGRAMS
    • G09B29/00Maps; Plans; Charts; Diagrams, e.g. route diagram
    • G09B29/003Maps
    • G09B29/004Map manufacture or repair; Tear or ink or water resistant maps; Long-life maps
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/38Electronic maps specially adapted for navigation; Updating thereof
    • G01C21/3804Creation or updating of map data
    • G01C21/3807Creation or updating of map data characterised by the type of data
    • G01C21/3811Point data, e.g. Point of Interest [POI]
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/38Electronic maps specially adapted for navigation; Updating thereof
    • G01C21/3804Creation or updating of map data
    • G01C21/3833Creation or updating of map data characterised by the source of data
    • G01C21/3848Data obtained from both position sensors and additional sensors
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T17/00Three dimensional [3D] modelling, e.g. data description of 3D objects
    • GPHYSICS
    • G09EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
    • G09BEDUCATIONAL OR DEMONSTRATION APPLIANCES; APPLIANCES FOR TEACHING, OR COMMUNICATING WITH, THE BLIND, DEAF OR MUTE; MODELS; PLANETARIA; GLOBES; MAPS; DIAGRAMS
    • G09B29/00Maps; Plans; Charts; Diagrams, e.g. route diagram
    • G09B29/003Maps
    • G09B29/005Map projections or methods associated specifically therewith
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2200/00Indexing scheme for image data processing or generation, in general
    • G06T2200/04Indexing scheme for image data processing or generation, in general involving 3D image data

Definitions

  • Embodiments of the present invention relate to a map construction technology, and in particular, to an indoor map construction method, apparatus, and storage medium.
  • the invention provides a method and a device for constructing an indoor map, so as to solve the problem that the staff capacity is high and the labor is large in the process of constructing the indoor map.
  • an embodiment of the present invention provides a method for constructing an indoor map.
  • the indoor map construction method includes:
  • the indoor three-dimensional map is cut to form an indoor map.
  • an embodiment of the present invention further provides an indoor map construction device.
  • the indoor map construction device includes:
  • a laser point cloud data acquisition module configured to acquire laser point cloud data collected by a laser point cloud device
  • a three-dimensional graph construction module configured to register the laser point cloud data based on key frame point cloud data included in the laser point cloud data, and form an indoor three-dimensional map according to the registration result;
  • An indoor map building module is configured to cut the indoor three-dimensional map to form an indoor map.
  • an embodiment of the present invention further provides a non-volatile computer storage medium, where the computer storage medium stores one or more modules, when the one or more modules are executed by an indoor map construction method.
  • the device is executed, the device is caused to perform the following operations:
  • the indoor three-dimensional map is cut to form an indoor map.
  • the indoor map construction method, device and storage medium register the acquired laser point cloud data based on the key frame point cloud data to form an indoor three-dimensional map, and perform the indoor three-dimensional map.
  • Cutting to form an indoor map can solve the current situation of building indoor maps in the absence of CAD drawings of buildings.
  • a large number of workers with professional drawing skills are required to personally survey the indoor environment, which requires high staff and labor.
  • the problem of large quantity has achieved the goal of reducing the staff's ability, reducing the workload of the staff, and improving the accuracy of the built indoor map.
  • Embodiment 1 is a flowchart of a method for constructing an indoor map according to Embodiment 1 of the present invention
  • FIG. 2 is a schematic diagram of a method for determining key frame point cloud data according to Embodiment 1 of the present invention
  • FIG. 3 is a flowchart of a method for constructing an indoor map according to Embodiment 2 of the present invention.
  • FIG. 5 is a schematic diagram of determining whether a moving track of a laser point cloud device constitutes a closed figure according to Embodiment 3 of the present invention.
  • FIG. 6 is a flowchart of a method for constructing an indoor map according to Embodiment 4 of the present invention.
  • FIG. 7 is a schematic structural diagram of an indoor map construction apparatus according to Embodiment 5 of the present invention.
  • FIG. 8 is a schematic diagram of a hardware structure of an apparatus for performing a map construction method according to Embodiment 7 of the present invention.
  • Embodiment 1 is a flowchart of a method for constructing an indoor map according to Embodiment 1 of the present invention.
  • the present embodiment is applicable to constructing an indoor map without a CAD drawing of a building, and the method can be executed by an indoor map construction device.
  • the device can be implemented by means of hardware and/or software.
  • the indoor map construction method specifically includes the following steps:
  • Laser point cloud data refers to the use of laser to obtain the spatial coordinates of each sampling point on the surface of the object under the same spatial reference system, and obtain a series of massive points that express the spatial distribution of the target and the characteristics of the target surface. set.
  • the laser point cloud data is collected by a laser point cloud device, which can be integrated into a staff backpack or a movable collection platform.
  • the laser point cloud device traverses the main road of the entire room as the staff moves, and in the above process, the laser point cloud device is set at intervals.
  • the laser point cloud data of the entire room is collected.
  • Registration refers to the process of converting laser point cloud data collected at different times into the same coordinate system. Specifically, in the process of the laser point cloud device traversing the main road of the entire room as the staff moves, the laser point cloud data collected for each frame is relative to the laser point cloud data acquisition moment laser point cloud In terms of the spatial coordinate system of the device, and the different acquisition moments, the spatial coordinate system of the laser point cloud device is different. In order to construct a three-dimensional map of the room to be built indoor map, it is necessary to reposition the laser point cloud data in different spatial coordinate systems to generate a three-dimensional map under a unified coordinate system, which is the registration of the laser point cloud data.
  • Key frame point cloud data refers to point cloud data used as a registration reference.
  • multiple methods can be used to determine key frame point cloud data.
  • the key frame point cloud data of the laser point cloud data corresponding to the scene may be determined according to the collection scene of the laser point cloud data; or, for each frame of the laser point cloud data, the previous frame laser point cloud data is determined as The key frame point cloud data of the frame laser point cloud data; or, according to the acquisition time of the laser point cloud data, the key frame point cloud data included in the laser point cloud data is determined.
  • the key frame point cloud data corresponding to the scene determined according to the collection scene of the laser point cloud data may be the first frame collected after the staff enters a certain room.
  • the laser point cloud data may also be the first frame or a certain frame of laser point cloud data collected after the acquisition scene is abrupt with the movement of the staff, or may be several frames of laser point cloud data collected under a certain scene.
  • the laser point cloud data obtained after registration is superimposed.
  • the worker carries the laser point cloud device to walk along the main road of the room clockwise from the point A1 (the dotted line in FIG. 2 indicates the path of walking), and collects To many frames of laser point cloud data.
  • An Point motion to point B1 the scene is abrupt, the laser point cloud data collected at point B1 can be used as key frame point cloud data, and the laser point cloud data collected from point B2 to point Bm is registered.
  • the key point cloud data included in the laser point cloud data is determined, which may be the laser point cloud data collected at the interval setting time, or may select several frames in the set time interval.
  • the laser point cloud data is registered and the laser point cloud data obtained after the superposition is obtained.
  • the acquisition time corresponding to the first frame laser point cloud data collected after entering a certain room is regarded as the 0 time, and the time is used as the starting time for calculating the key frame point cloud data.
  • the laser point cloud data collected every 3 s is taken as key frame point cloud data.
  • the laser point cloud data collected in 0-3s is registered based on the laser point cloud data collected at time 0; the laser point cloud data collected in 3-6s is based on the laser point collected in the 3s Cloud data is registered; and so on.
  • a part or a part of the features in the laser point cloud data may be registered (ie, coarse registration), or All points in the laser point cloud data are registered (ie, fine registration).
  • some of the features in the laser point cloud data mentioned herein include, but are not limited to, one or more of the following features: normal vector direction, curvature, and histogram.
  • ICP algorithm the nearest point iterative algorithm
  • the laser point cloud data may be registered or multiple times based on key frame point cloud data, and the indoor three-dimensional map is formed according to the last registration result. Further, in each registration process, only partial or partial features in the laser point cloud data may be registered; all points in the laser point cloud data may also be registered; Partial or partial features in the cloud data are registered, and then all points in the laser point cloud data are registered.
  • the laser point cloud data is first registered; secondly, based on the second key frame point cloud data included in the laser point cloud data, the second registration of the laser point cloud data after the first registration is performed
  • the second key frame point cloud data is different from the first key frame point cloud data; finally, an indoor three-dimensional map is formed according to the second registration result.
  • the first key frame point cloud data may be the previous frame laser point cloud data for the laser point cloud data to be registered
  • the second key frame point cloud data may be the collection scene according to the laser point cloud data.
  • the determined laser point cloud data corresponding to the scene may also be the key frame point cloud data included in the laser point cloud data determined according to the collection time of the laser point cloud data.
  • the method can effectively improve the success rate of laser point cloud data registration by two registrations, and can reduce the cumulative error in the registration process.
  • a library indoor map is mainly used to help users quickly find a specific numbered bookshelf in the borrowing area.
  • the indoor map needs to include the bookshelf in the borrowing area. Location information.
  • the indoor map includes only the position information of the bookshelf except for the location information of the wall, window and door of the borrowing area, and does not include the reader who is located in the borrowing area when the laser point cloud data is collected.
  • the indoor map only needs to include the location information of the wall, the window and the door.
  • the indoor space is divided into different indoor areas according to a preset partitioning rule; and the indoor three-dimensional map is determined according to the regional characteristics of the indoor area. Cutting is performed to form an indoor map, the area features including height.
  • the preset partition rules here include the height of the indoor floor to the roof. Illustratively, the height of the floor to the roof of a room is 2.5m, the stage is set in the middle of the room, and the stage is 0.5m above the ground.
  • the room can be divided into two areas according to the height of the ground to the roof, the first area is an area with a stage, and the second area is an unstaged Area.
  • the height of the person is up to 1.8m.
  • the distance from the ground is 2.4m, and the three-dimensional map of the area is cut.
  • the distance from the ground is 2m.
  • the three-dimensional map of the area is cut.
  • the cross-sectional view after cutting the three-dimensional map of the first region is combined with the cross-sectional view after the three-dimensional map of the second region is cut to form an indoor map of the room.
  • the acquired laser point cloud data is registered based on the key frame point cloud data to form an indoor three-dimensional map, and the formed indoor three-dimensional map is cut to form an indoor map, which can solve the current no-building.
  • the CAD drawing of the object in the process of constructing the indoor map, a large number of staff with professional drawing ability are required to personally survey the indoor environment, and the problem of high staff capacity and labor volume is realized, and the staff is reduced. The ability to reduce the amount of labor of the staff and improve the accuracy of the built indoor map.
  • the obtained indoor map can also be projected onto the corresponding outdoor map to generate a plane raster map of the corresponding position. Further, in order to meet the needs of different users for the map, the planar raster map can also be processed to form a vector map.
  • FIG. 3 is a flowchart of a method for constructing an indoor map according to Embodiment 2 of the present invention.
  • the present embodiment is based on key frame point cloud data included in the laser point cloud data.
  • the laser point cloud data is further added with a feature: acquiring an inertial measurement unit attitude acquired by the inertial measurement unit, wherein the inertial measurement unit is rigidly connected to the laser point cloud device; and correcting according to the posture of the inertial measurement unit The laser point cloud data.
  • S210 Acquire an inertial measurement unit attitude acquired by an inertial measurement unit, wherein the inertial measurement unit is rigidly connected to the laser point cloud device;
  • An Inertial Measurement Unit is a device that measures the three-axis attitude angle (or angular rate) and acceleration of an object.
  • the inertial measurement unit is rigidly connected to the laser point cloud device, and is integrated into the staff's backpack or the movable collection platform. As the staff moves, the main road of the entire room is traversed, and the entire room is collected in the laser point cloud device.
  • the laser point cloud data is acquired while the inertial measurement unit poses. It should be noted that when the inertial measurement unit is used to acquire its own attitude, it is necessary to clarify the corresponding relationship between the acquired inertial measurement unit attitude and the collected laser point cloud data, that is, the inertial measurement unit and the laser point cloud are realized.
  • the device is synchronized.
  • the attitude of the laser point cloud device can be obtained by combining the attitude of the inertial measurement unit and the positional relationship between the inertial measurement unit and the laser point cloud device. That is, there is a one-to-one correspondence between the attitude of the inertial measurement unit and the attitude of the laser point cloud device.
  • the specific implementation method of the step is to determine the time relationship between the laser point cloud data and the inertial measurement unit attitude time, and the corresponding relationship between the attitude of the inertial measurement unit and the attitude space of the laser point cloud device, and determine the time when the laser point cloud data is collected.
  • the attitude of the laser point cloud device and then correct the laser point cloud data according to the attitude of the laser point cloud device at the time of laser point cloud data acquisition.
  • all the laser point cloud data collected may be corrected when the laser point cloud data is corrected, or only part of the laser point cloud data may be corrected. For example, only the laser point cloud data to be registered may be corrected.
  • the laser point cloud data is corrected by using the inertial measurement unit posture acquired by the inertial measurement unit, which can effectively reduce the measurement error of the laser point cloud data, thereby improving the accuracy of the constructed indoor map.
  • FIG. 4 is a flowchart of a method for constructing an indoor map according to Embodiment 3 of the present invention.
  • the present embodiment will be based on key frame point cloud data included in the laser point cloud data. Aligning the laser point cloud data and forming an indoor three-dimensional map according to the registration result is optimized to: match the laser point cloud data based on the key frame point cloud data included in the laser point cloud data Correctly determining a movement trajectory of the laser point cloud device during the process of acquiring the laser point cloud data according to the registration result; determining whether the movement trajectory constitutes a closed pattern, and if the moving trajectory constitutes a closed pattern, using a loopback The optimization algorithm corrects the laser point cloud data; and combines the corrected laser point cloud data to form an indoor three-dimensional map.
  • the step is specifically based on the obtained rotation matrix of each frame of the laser point cloud data at the time of registration, and the laser point cloud device is launched at the time of collecting the laser point cloud data, relative to the time of collecting the key frame point cloud data.
  • the position change information is obtained according to the position change information, and the specific position of the laser point cloud device is obtained at the time of collecting the laser point cloud data of each frame, thereby obtaining the movement track of the laser point cloud device.
  • determining whether the moving track constitutes a closed figure does not mean that the moving track constitutes a strict head-to-tail coincidence pattern, but refers to whether the moving track within the range allowed by the error constitutes a basic coincidence of the head and the tail. Graphics.
  • determining whether the movement trajectory constitutes a closed figure there are various methods specifically for determining whether the movement trajectory constitutes a closed figure.
  • determining whether the distance between the specific location of the laser point cloud device at each laser spot cloud data acquisition time and the position of the laser point cloud device at the time of acquiring the first frame laser point cloud data in the room during registration is less than or Is equal to a certain preset value. For example, as shown in FIG. 5, after the worker carries the laser point cloud device into the room 1, a circle is traveled along the main road of the room in a clockwise direction from point A (the dotted line in FIG. 5 is the walking trajectory of the worker), Reach point C.
  • 0.5m is selected as the preset value for judging whether the moving track constitutes a closed figure in the process of judging, and the distance value between the point A and the point C is determined and the preset value (0.5m) is determined.
  • the size relationship If the distance between point A and point C is greater than 0.5 m, it indicates that the movement trajectory does not constitute a closed pattern; if the distance value between point A and point C is less than or equal to 0.5 m, the movement trajectory constitutes a closed figure. .
  • the laser point cloud data is corrected by using a loopback optimization algorithm, specifically, as shown in FIG. 5, if the distance between the point A and the point C is indeed less than or equal to 0.5 m. , the movement trajectory constitutes a closed figure.
  • the laser point cloud data collected at point C and the laser point cloud data collected at point A are forcibly combined to obtain an error matrix of laser point cloud data collected at point C.
  • the error matrix of the laser point cloud data collected at point C is evenly distributed to each frame of laser point cloud data collected during the process of moving from point A to point C, and the increment corresponding to each frame of laser point cloud data is obtained.
  • the incremental matrix is the error that should be subtracted for each frame of laser point cloud data.
  • correcting the laser point cloud data refers to subtracting the incremental matrix from each frame of laser point cloud data collected during the process of moving from point A to point C.
  • the present embodiment determines whether the moving trajectory of the laser point cloud data constitutes a closed graph, and uses the loopback optimization algorithm to correct all laser point cloud data capable of forming a closed graph. Effectively removing the cumulative error of the formed three-dimensional map can achieve the purpose of further improving the accuracy of the indoor map.
  • FIG. 6 is a flowchart of a method for constructing an indoor map according to Embodiment 4 of the present invention. This embodiment further adds a filtering processing step based on the above embodiments.
  • S430 Filtering noise interference in the indoor three-dimensional image, small-sized object interference, or dynamic interference caused by continuous changes in the position of the person or the object.
  • the one or more steps S410, S420, and S430 may be specifically selected for filtering processing according to specific situations.
  • the required filtering algorithms may be the same or different.
  • a corresponding filtering algorithm is selected according to the type of interference for filtering. For example, for noise interference, the outlier removal method can be selected to filter the laser point cloud data.
  • the filtering processing step by adding the filtering processing step, the problem that the formed indoor map is low due to the existence of the interference can be effectively solved, and the purpose of further improving the accuracy of the constructed indoor map can be achieved.
  • Embodiments 1 through 4 of the present application provide various methods for improving the accuracy of the constructed indoor map from various angles. Each method does not affect each other during the execution process and can be superimposed on each other.
  • FIG. 7 is an indoor map construction apparatus according to Embodiment 5 of the present invention.
  • the indoor map construction device includes a laser point cloud data acquisition module 510, a three-dimensional map construction module 520, and an indoor map construction module 530.
  • the laser point cloud data acquiring module 510 is configured to acquire laser point cloud data collected by the laser point cloud device;
  • the three-dimensional graph construction module 520 is configured to register the laser point cloud data based on the key frame point cloud data included in the laser point cloud data, and form an indoor three-dimensional map according to the registration result;
  • the indoor map construction module 530 is configured to cut the indoor three-dimensional map to form an indoor map.
  • the acquired laser point cloud data is registered based on the key frame point cloud data to form an indoor three-dimensional map, and the formed indoor three-dimensional map is cut to form an indoor map, which can solve the current no-building.
  • CAD drawings in the process of constructing indoor maps, a large number of staff with professional drawing skills are required to personally survey the indoor environment, and the problem of high staff capacity and labor volume is achieved, and the ability of the staff is reduced. It is required to reduce the workload of the staff and improve the accuracy of the constructed indoor map.
  • the indoor map construction device may further include:
  • a laser point cloud correction module configured to acquire an inertial measurement unit posture acquired by the inertial measurement unit before the laser point cloud data is registered based on the key frame point cloud data included in the laser point cloud data, where The inertial measurement unit is rigidly connected to the laser point cloud device; and the laser point cloud data is corrected according to the posture of the inertial measurement unit.
  • the three-dimensional graph construction module 520 may further include:
  • a registration unit configured to register the laser point cloud data based on key frame point cloud data included in the laser point cloud data
  • a movement trajectory determining unit configured to determine, according to the registration result, a movement trajectory of the laser point cloud device during the process of acquiring the laser point cloud data
  • a laser point cloud data correcting unit configured to determine whether the moving track constitutes a closed figure, and if the moving track constitutes a closed figure, correcting the laser point cloud data by using a loopback optimization algorithm
  • the first three-dimensional map forming unit is configured to flatten the corrected laser point cloud data to form an indoor three-dimensional map.
  • the three-dimensional graph construction module 520 can include:
  • a first registration unit configured to perform the first registration on the laser point cloud data based on the first key frame point cloud data included in the laser point cloud data
  • a second registration unit configured to perform a second registration on the first registered laser point cloud data based on the second key frame point cloud data included in the laser point cloud data, wherein the second key Frame point cloud According to the first key frame point cloud data;
  • a second three-dimensional map forming unit configured to form an indoor three-dimensional map according to the second registration result.
  • the previous frame laser point cloud data is determined as the key frame point cloud data of the frame laser point cloud data.
  • the indoor map construction device may further include:
  • a first filtering module configured to perform specular interference and/or miscellaneous inclusion of the laser point cloud data before the laser point cloud data is registered based on key frame point cloud data included in the laser point cloud data Point interference is filtered.
  • the three-dimensional graph construction module 520 is specifically configured to register the laser point cloud data based on the key frame point cloud data included in the laser point cloud data, and in the registration process, the person or the object Dynamic disturbance caused by continuous change of position is filtered; according to the registration result and the filtering result, the laser point cloud data is combined to form an indoor three-dimensional map.
  • the indoor map construction device may further include:
  • a second filtering module configured to perform noise interference, small-sized object interference, or dynamic interference caused by continuous changes of a person or an object position in the indoor three-dimensional image before cutting the indoor three-dimensional image to form an indoor map Filtering is performed.
  • the indoor map construction module 530 includes:
  • the indoor space dividing unit is configured to divide the indoor space into different indoor areas according to a preset partitioning rule
  • the three-dimensional image cutting unit is configured to cut the indoor three-dimensional map according to the regional feature of the indoor region to form an indoor map, and the regional feature includes a height.
  • the above product can perform the method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.
  • a non-volatile computer storage medium storing one or more modules that, when executed by a device executing an indoor map construction method, cause the device to perform the following operations :
  • the indoor three-dimensional map is cut to form an indoor map.
  • the method may further include:
  • the laser point cloud data is corrected according to the inertial measurement unit attitude.
  • the laser point cloud data is registered based on key frame point cloud data included in the laser point cloud data, and an indoor three-dimensional map is formed according to the registration result.
  • the corrected laser point cloud data is combined to form an indoor three-dimensional map.
  • the laser point cloud data is registered based on key frame point cloud data included in the laser point cloud data, and an indoor three-dimensional map is formed according to the registration result.
  • the key frame point cloud data can be obtained as follows:
  • the previous frame laser point cloud data is determined as the key frame point cloud data of the frame laser point cloud data.
  • the method may further include:
  • the laser point cloud data is registered based on key frame point cloud data included in the laser point cloud data, and an indoor three-dimensional map is formed according to the registration result.
  • mapping according to the key frame point cloud data included in the laser point cloud data, the laser point cloud data, and filtering dynamic interference caused by continuous changes of the position of the person or the object during the registration process;
  • the laser point cloud data are combined to form an indoor three-dimensional map.
  • the method may further include:
  • Filtering is caused by noise interference in the indoor three-dimensional map, small-sized object interference, or dynamic interference due to continuous changes in the position of the person or object.
  • the indoor three-dimensional map is cut. Cut into an indoor map, which can include:
  • the indoor three-dimensional map is cut to form an indoor map according to the regional characteristics of the indoor area, and the regional features include a height.
  • FIG. 8 is a schematic structural diagram of hardware of an apparatus for performing an indoor map construction method according to Embodiment 7 of the present invention.
  • the device includes:
  • One or more processors 610, one processor 610 is taken as an example in FIG. 8;
  • Memory 620 and one or more modules.
  • the device may also include an input device 630 and an output device 640.
  • the processor 610, the memory 620, the input device 630, and the output device 640 in the device may be connected by a bus or other means, and the bus connection is taken as an example in FIG.
  • the memory 620 is used as a computer readable storage medium for storing software programs, computer executable programs, and modules, such as program instructions/modules corresponding to the indoor map construction method in the embodiment of the present invention (for example, as shown in FIG. 7)
  • the processor 610 executes various functional applications and data processing of the server by executing software programs, instructions, and modules stored in the memory 620, that is, implementing the indoor map construction method in the above method embodiments.
  • the memory 620 may include a storage program area and an storage data area, wherein the storage program area may store an operating system, an application required for at least one function; the storage data area may store data created according to usage of the terminal device, and the like.
  • memory 620 can include high speed random access memory, and can also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device.
  • memory 620 can further include memory remotely located relative to processor 610, which can be connected to the terminal device over a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.
  • Input device 630 can be configured to receive input numeric or character information, and to generate a user with the terminal Set and signal input related to function control.
  • the output device 640 can include a display device such as a display screen.
  • the one or more modules are stored in the memory 620, and when executed by the one or more processors 610, perform the following operations:
  • the indoor three-dimensional map is cut to form an indoor map.
  • the method before registering the laser point cloud data based on the key frame point cloud data included in the laser point cloud data, the method further includes:
  • the laser point cloud data is corrected according to the inertial measurement unit attitude.
  • registering the laser point cloud data, and forming an indoor three-dimensional map according to the registration result including:
  • the corrected laser point cloud data is combined to form an indoor three-dimensional map.
  • registering the laser point cloud data, and forming an indoor three-dimensional map according to the registration result including:
  • the previous frame laser point cloud data is determined as the key frame point cloud data of the frame laser point cloud data.
  • the method before registering the laser point cloud data based on the key frame point cloud data included in the laser point cloud data, the method further includes:
  • registering the laser point cloud data, and forming an indoor three-dimensional map according to the registration result including:
  • mapping according to the key frame point cloud data included in the laser point cloud data, the laser point cloud data, and filtering dynamic interference caused by continuous changes of the position of the person or the object during the registration process;
  • the laser point cloud data are combined to form an indoor three-dimensional map.
  • the method further includes:
  • Filtering is caused by noise interference in the indoor three-dimensional map, small-sized object interference, or dynamic interference due to continuous changes in the position of the person or object.
  • cutting the indoor three-dimensional image to form an indoor map including:
  • the indoor three-dimensional map is cut to form an indoor map according to the regional characteristics of the indoor area, and the regional features include a height.
  • the present invention can be implemented by software and necessary general hardware, and can also be implemented by hardware, but in many cases, the former is a better implementation. .
  • the technical solution of the present invention which is essential or contributes to the prior art, can be embodied in the form of a software product.
  • the software product can be stored in a computer readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc.
  • a number of instructions are included to cause a computer device (which may be a personal computer, server, or network device, etc.) to perform the methods described in various embodiments of the present invention.
  • the units and modules included in the embodiment are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be implemented;
  • the specific names of the respective functional units are also for convenience of distinguishing from each other and are not intended to limit the scope of protection of the present invention.

Abstract

一种室内地图构建方法、装置和存储介质。室内地图构建方法包括:获取激光点云设备采集的激光点云数据(S110);基于激光点云数据包含的关键帧点云数据,对激光点云数据进行配准,并依据配准结果形成室内三维图(S120);对室内三维图进行切割形成室内地图(S130)。室内地图构建方法可以解决目前在无建筑物的CAD图纸的情况下,构建室内地图的过程中,需要大量具备专业绘图能力的工作人员亲自对室内环境进行测绘,对工作人员能力要求高且劳动量大的问题,实现降低对工作人员能力的要求,减小工作人员的劳动量,提高所构建的室内地图的精度的目的。

Description

一种室内地图构建方法、装置和存储介质
本专利申请要求于2016年3月31日提交的、申请号为201610202360.9、申请人为百度在线网络技术(北京)有限公司、发明名称为“一种室内地图构建方法及装置”的中国专利申请的优先权,该申请的全文以引用的方式并入本申请中。
技术领域
本发明实施例涉及地图构建技术,尤其涉及一种室内地图构建方法、装置和存储介质。
背景技术
随着计算机技术、无线定位技术、地球信息系统及移动互联技术的飞速发展,基于位置的服务成为现实并在实际中得到了广泛应用。在室内环境中,如机场大厅、展厅、仓库、超市、图书馆、地下停车场、矿井等环境中,常常需要确定移动终端或其持有者、设施与物品在室内的位置信息,并提供相应的附加诸如导航,搜索查询等基于室内位置的应用服务。然而,由于室内建筑数量巨大且室内环境复杂多变,如超市、展厅装修布局的周期性改变,对室内位置服务的地图更新的时效性提出了严峻的挑战。
目前,在构建室内地图的过程中,若没有待构建室内地图的建筑物的原始CAD图纸的情况下,需要工作人员对室内各房间墙壁、门窗等尺寸进行测量,进而得到建筑的室内地图。这种室内地图的构建过程需要大量具备专业绘图能力的工作人员亲自对室内环境进行测绘,对工作人员能力要求高且劳动量大。
发明内容
本发明提供了一种室内地图构建方法及装置,以解决目前构建室内地图的过程中对工作人员能力要求高且劳动量大的问题。
第一方面,本发明实施例提供了一种室内地图构建方法。该室内地图构建方法包括:
获取激光点云设备采集的激光点云数据;
基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图;
对所述室内三维图进行切割形成室内地图。
第二方面,本发明实施例还提供了一种室内地图构建装置。该室内地图构建装置包括:
激光点云数据获取模块,用于获取激光点云设备采集的激光点云数据;
三维图构建模块,用于基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图;
室内地图构建模块,用于对所述室内三维图进行切割形成室内地图。
第三方面,本发明实施例还提供了一种非易失性计算机存储介质,所述计算机存储介质存储有一个或者多个模块,当所述一个或者多个模块被一个执行室内地图构建方法的设备执行时,使得所述设备执行如下操作:
获取激光点云设备采集的激光点云数据;
基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图;
对所述室内三维图进行切割形成室内地图。
本发明实施例提供的室内地图构建方法、装置和存储介质,通过基于关键帧点云数据,对所获取的激光点云数据进行配准,形成室内三维图,并对所形成的室内三维图进行切割形成室内地图,可以解决目前在无建筑物的CAD图纸的情况下,构建室内地图的过程中,需要大量具备专业绘图能力的工作人员亲自对室内环境进行测绘,对工作人员能力要求高且劳动量大的问题,实现了降低对工作人员能力的要求,减小工作人员的劳动量,提高所构建的室内地图的精度的目的。
附图说明
图1是本发明实施例一提供的一种室内地图构建方法的流程图;
图2是本发明实施例一提供的一种关键帧点云数据确定方法的原理图;
图3是本发明实施例二提供的一种室内地图构建方法的流程图;
图4是本发明实施例三提供的一种室内地图构建方法的流程图;
图5是本发明实施例三提供的一种判断激光点云设备的移动轨迹是否构成闭合图形的原理图;
图6是本发明实施例四提供的一种室内地图构建方法的流程图;
图7是本发明实施例五提供的一种室内地图构建装置的结构示意图。
图8是本发明实施例七提供的一种执行地图构建方法的设备的硬件结构示意图。
具体实施方式
下面结合附图和实施例对本发明作进一步的详细说明。可以理解的是,此处所描述的具体实施例仅仅用于解释本发明,而非对本发明的限定。另外还需要说明的是,为了便于描述,附图中仅示出了与本发明相关的部分而非全部结构。
实施例一
图1为本发明实施例一提供的一种室内地图构建方法的流程图,本实施例可适用于在无建筑物CAD图纸的情况下,构建室内地图,该方法可以由室内地图构建装置来执行,该装置可通过硬件和/或软件的方式实现。
该室内地图构建方法具体包括如下步骤:
S110、获取激光点云设备采集的激光点云数据。
激光点云数据,是指利用激光在同一空间参考系下获取物体表面每个采样点的空间坐标,得到的是一系列表达目标空间分布和目标表面特性的海量点的 集合。
该激光点云数据通过激光点云设备进行采集,该激光点云设备可以集成于工作人员的背包或者可移动的采集平台上。在需要对待构建室内地图的房间的激光点云数据进行采集时,该激光点云设备随着工作人员的移动,遍历整个房间的主干道,并在上述过程中,激光点云设备间隔设定时间对整个房间的激光点云数据进行采集。
S120、基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图。
配准是指将不同时刻采集的激光点云数据转换至同一坐标系下的过程。具体而言,在激光点云设备随着工作人员的移动,遍历整个房间的主干道的过程中,所采集的每一帧激光点云数据都是相对于该激光点云数据采集时刻激光点云设备的空间坐标系而言的,并且不同采集时刻,激光点云设备的空间坐标系不同。为了构建待构建室内地图的房间的三维图,需要将不同空间坐标系下的激光点云数据进行重新定位,生成一个统一坐标系下的三维图,这就是激光点云数据的配准。
关键帧点云数据是指用做配准基准的点云数据,在具体进行配准时可以采用多种方法确定关键帧点云数据。例如,可以依据激光点云数据的采集场景,确定与该场景对应的激光点云数据的关键帧点云数据;或者,对于每一帧激光点云数据,将前一帧激光点云数据确定为该帧激光点云数据的关键帧点云数据;或者,依据激光点云数据的采集时间,确定激光点云数据包含的关键帧点云数据。
在上述关键帧点云数据的确定方法中,依据激光点云数据的采集场景,确定的与该场景对应的关键帧点云数据,具体可以是工作人员进入到某房间后所采集的第一帧激光点云数据,也可以是随着工作人员的移动,采集场景发生突变后所采集的第一帧或某一帧激光点云数据,还可以是将某场景下采集的几帧激光点云数据配准叠合后得到的激光点云数据。示例性地,如图2所示,在房间1内,工作人员携带激光点云设备从A1点开始沿顺时针方向沿房间主干道行走一圈(图2中虚线表示其行走的路径),采集到众多帧激光点云数据。由An 点运动到B1点,场景发生突变,可以将B1点所采集到的激光点云数据作为关键帧点云数据,对B2点至Bm点所采集到的激光点云数据进行配准。
依据激光点云数据的采集时间,确定的激光点云数据包含的关键帧点云数据,具体可以是间隔设定时间所采集的激光点云数据,还可以是在设定时间间隔内选取几帧激光点云数据配准叠合后得到的激光点云数据。示例性地,将进入某房间后所采集到的第一帧激光点云数据对应的采集时刻视作为0时刻,并将该时刻作为计算关键帧点云数据的起始时刻。将每隔3s所采集的激光点云数据作为关键帧点云数据。在配准时,0-3s内所采集的激光点云数据,基于0时刻所采集的激光点云数据进行配准;3-6s内所采集的激光点云数据,基于第3s所采集的激光点云数据进行配准;依次类推。
需要说明的是,在确定关键帧点云数据时,依据激光点云数据的采集场景,确定与该场景对应的激光点云数据的关键帧点云数据;以及依据激光点云数据的采集时间,确定激光点云数据包含的关键帧点云数据这两种方法,容易出现配准失败,但是在配准过程中累积误差小。而采用对于每一帧激光点云数据,将前一帧激光点云数据确定为该帧激光点云数据的关键帧点云数据的方法,虽然配准成功率高,但是累积误差大。
基于所述激光点云数据包含的关键帧点云数据,对激光点云数据进行配准时,可以对该激光点云数据中的部分点或部分特征进行配准(即粗配准),也可以对该激光点云数据中全部点进行配准(即精配准)。需要说明的是,这里所提到的激光点云数据中的部分特征包括但不限于下述特征的一种或多种:法向量方向、曲率以及直方图。若对激光点云数据中全部点进行配准,可以采用最近点迭代算法(ICP算法)进行配准。在具体配准的过程中,可以基于关键帧点云数据对激光点云数据进行一次配准或多次配准,并根据最后一次配准结果拼合形成室内三维图。进一步地,在每一次配准的过程中,可以仅对激光点云数据中的部分点或部分特征进行配准;也可以对激光点云数据中全部点进行配准;还可以先对激光点云数据中的部分点或部分特征进行配准,然后对激光点云数据中全部点进行配准。
示例性地,首先,基于所述激光点云数据包含的第一关键帧点云数据,对 所述激光点云数据进行第一次配准;其次,基于所述激光点云数据包含的第二关键帧点云数据,对第一次配准后的激光点云数据进行第二次配准,其中所述第二关键帧点云数据与所述第一关键帧点云数据不同;最后,根据第二次配准结果,形成室内三维图。典型地,这里第一关键帧点云数据可以为针对待配准激光点云数据的前一帧激光点云数据,第二关键帧点云数据可以为依据所述激光点云数据的采集场景,所确定的与该场景对应的激光点云数据;也可以为依据所述激光点云数据的采集时间,所确定的所述激光点云数据包含的关键帧点云数据。本方法通过两次配准可以既有效提高激光点云数据配准的成功率,又能够降低配准过程中的累积误差。
S130、对所述室内三维图进行切割形成室内地图。
在对室内三维图进行切割形成室内地图的过程中,可以根据室内地图的具体用途,选择不同的切割标准进行切割。例如,某图书馆室内地图主要用于帮助用户快速找到借阅区某特定编号的书架,该室内地图中除了需要包括该借阅区墙壁、窗户以及门的位置信息外,还需要包括借阅区内各书架的位置信息。在对该室内三维图进行切割时,考虑到书架的高度通常为2m,人的高度最高为1.8m,可选取距离地面1.9m处,对室内三维图进行切割形成室内地图,这样可以确保所形成的室内地图中除包括该借阅区墙壁、窗户以及门的位置信息外,仅包括书架的位置信息,不包括激光点云数据采集时刻位于借阅区的阅读者。又例如,若某图书馆室内地图主要用于帮助用户定位并迅速找到出口,该室内地图仅需要包括墙壁、窗户以及门的位置信息即可。在对该室内三维图进行切割时,考虑到该图书馆内书架的高度通常为2m,屋顶到地面的高度为3m,可选取距离地面2.5m处,对室内三维图进行切割形成室内地图。这样可以确保所形成的室内地图中仅包括墙壁、窗户以及门的位置信息。
进一步地,考虑到对于同一室内区域具有不同的区域特征,可选地,根据预设的分区规则,将室内空间划分为不同室内区域;依据所述室内区域的区域特征,对所述室内三维图进行切割形成室内地图,所述区域特征包括高度。这里预设的分区规则包括室内地面到屋顶的高度。示例性地,某房间地面到屋顶的高度为2.5m,房间中间位置设置有舞台,舞台高于地面0.5m,若需要构建该 房间仅包含墙壁、窗户以及门的位置信息的室内地图时,可以根据地面到屋顶的高度,将该房间分为两个区域,第一区域为设置有舞台的区域,第二区域为未设舞台的区域。在考虑到采集点云数据过程中,人的高度最高为1.8m,可以针对第一区域,选取距地面2.4m,对该区域的三维图进行切割;针对第二区域,选取距地面2m,对该区域的三维图进行切割。将第一区域三维图切割后的截面图与第二区域三维图切割后的截面图拼合,形成房间的室内地图。通过结合室内各区域的区域特征,对各区域的三维图进行分区域切割后拼合形成室内地图,这样可以更好地满足用户对是室内地图的需求,提高用户体验。
本实施例技术方案通过基于关键帧点云数据,对所获取的激光点云数据进行配准,形成室内三维图,并对所形成的室内三维图进行切割形成室内地图,可以解决目前在无建筑物的CAD图纸的情况下,构建室内地图的过程中,需要大量具备专业绘图能力的工作人员亲自对室内环境进行测绘,对工作人员能力要求高且劳动量大的问题,实现了降低对工作人员能力的要求,减小工作人员的劳动量,提高所构建的室内地图的精度的目的。
在上述技术方案的基础上,还可以将所得到的室内地图投影到对应的室外地图上,生成相应位置的平面栅格地图。进一步地,为了满足不同用户对地图的需求,还可以对该平面栅格地图进行加工,形成矢量地图。
实施例二
图3为本发明实施例二提供的一种室内地图构建方法的流程图,本实施例在实施例一的基础上,在基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准之前进一步增加了特征:获取惯性测量单元采集到的惯性测量单元姿态,其中所述惯性测量单元与所述激光点云设备刚性连接;依据所述惯性测量单元姿态,修正所述激光点云数据。
本实施例提供的室内地图构建方法具体包括如下步骤:
S110、获取激光点云设备采集的激光点云数据。
S210、获取惯性测量单元采集到的惯性测量单元姿态,其中所述惯性测量单元与所述激光点云设备刚性连接;
惯性测量单元(Inertial measurement unit,IMU)是测量物体三轴姿态角(或角速率)以及加速度的装置。该惯性测量单元与激光点云设备刚性连接,同时集成于工作人员的背包或者可移动的采集平台上,随着工作人员的移动,遍历整个房间的主干道,并在激光点云设备采集整个房间的激光点云数据的同时采集惯性测量单元姿态。需要说明的是,利用惯性测量单元对其自身的姿态进行采集时,需要明确所采集的惯性测量单元姿态与所采集的激光点云数据时间上的对应关系,即实现惯性测量单元与激光点云设备同步采集。
S220、依据所述惯性测量单元姿态,修正所述激光点云数据。
由于惯性测量单元与激光点云设备刚性连接,可结合惯性测量单元姿态以及惯性测量单元与激光点云设备的位置关系得到激光点云设备的姿态。即惯性测量单元姿态与激光点云设备姿态在空间上存在一一对应的关系。
本步骤的具体实现方法为,依据激光点云数据与惯性测量单元姿态时间上对应关系,以及惯性测量单元姿态与激光点云设备姿态空间上的对应关系,确定在采集各激光点云数据时刻,激光点云设备的姿态,进而根据激光点云数据采集时刻激光点云设备的姿态,修正激光点云数据。
需要说明的,在修正激光点云数据时可以对所采集的全部激光点云数据进行修正,也可以仅修正部分激光点云数据,示例性地,仅修正需要配准的激光点云数据。
S120、基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图。
S130、对所述室内三维图进行切割形成室内地图。
本实施例利用惯性测量单元采集的惯性测量单元姿态修正所述激光点云数据,可以有效减小激光点云数据的测量误差,进而提高所构建的室内地图的精度。
实施例三
图4为本发明实施例三提供的一种室内地图构建方法的流程图,本实施例在上述各实施例的基础上,将基于所述激光点云数据包含的关键帧点云数据, 对所述激光点云数据进行配准,并依据配准结果形成室内三维图这一特征优化为:基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准;根据配准结果确定在采集所述激光点云数据的过程中所述激光点云设备的移动轨迹;确定所述移动轨迹是否构成闭合图形,若所述移动轨迹构成闭合图形,则利用回环优化算法,校正所述激光点云数据;拼合校正后的激光点云数据,形成室内三维图。
本实施例提供的室内地图构建方法具体包括如下步骤:
S110、获取激光点云设备采集的激光点云数据。
S310、基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准。
S320、根据配准结果确定在采集所述激光点云数据的过程中所述激光点云设备的移动轨迹。
本步骤具体为根据所配准的每一帧激光点云数据在配准时的得到的旋转矩阵,反推出在采集该激光点云数据时刻,相对于采集关键帧点云数据时刻,激光点云设备的位置变换信息,根据该位置变换信息得到在采集每一帧激光点云数据时刻,激光点云设备的具体位置,进而得到激光点云设备的移动轨迹。
S330、确定所述移动轨迹是否构成闭合图形,若所述移动轨迹构成闭合图形,则利用回环优化算法,校正所述激光点云数据。
需要说明的是,在本步骤中,确定所述移动轨迹是否构成闭合图形并非指判断所述移动轨迹构成严格的首尾重合的图形,而是指误差允许的范围之内移动轨迹是否构成首尾基本重合的图形。
具体用于确定所述移动轨迹是否构成闭合图形的方法有多种。示例性地,判断配准时每一帧激光点云数据采集时刻激光点云设备的具体位置与该激光点云设备在该室内采集第一帧激光点云数据时刻的位置之间的距离是否小于或等于某一预设值。例如,图5所示,工作人员携带激光点云设备进入到房间1后,从A点开始沿顺时针方向沿该房间的主干道行走一圈(图5中虚线为工作人员的行走轨迹),达到C点。假设在判断的过程中选取0.5m作为用于判断移动轨迹是否构成闭合图形的预设值,判断A点与C点之间的距离值与预设值(0.5m) 的大小关系。若A点与C点之间的距离值大于0.5m,则说明该移动轨迹未构成闭合图形;若A点与C点之间的距离值小于或等于0.5m,则说明该移动轨迹构成闭合图形。
若所述移动轨迹构成闭合图形,则利用回环优化算法,校正所述激光点云数据,具体是指,如图5所示,若A点与C点之间的距离值确实小于或等于0.5m,则说明该移动轨迹构成闭合图形,此时将C点所采集的激光点云数据与A点所采集的激光点云数据强行拼合在一起,得到C点所采集的激光点云数据的误差矩阵。将C点所采集的激光点云数据的误差矩阵平均分摊到从A点移动到C点这个过程中所采集的每一帧激光点云数据上,得到每一帧激光点云数据对应的增量矩阵,该增量矩阵即为每一帧激光点云数据应当扣除的误差。在本具体示例中,校正所述激光点云数据是指从A点移动到C点这个过程中所采集的每一帧激光点云数据中扣除该增量矩阵。
S340、拼合校正后的激光点云数据,形成室内三维图。
S130、对所述室内三维图进行切割形成室内地图。
在对所述激光点云数据进行配准,并依据配准结果形成室内三维图的过程中,由于每一帧激光点云数据在采集的过程中都存在误差,当每一帧激光点云数据在配准的过程中,会对各帧激光点云数据存在的误差进行累积。本实施例通过在对激光点云数据进行配准的过程中,判断激光点云数据的移动轨迹是否构成闭合图形,并利用回环优化算法对能够构成闭合图形的所有激光点云数据进行校正,可以有效去除所形成的三维图的累积误差,可达到进一步提高室内地图精度的目的。
实施例四
在利用本发明上述各实施例提供的室内地图构建方法构建室内地图的过程中,会存在很多干扰。例如,由于室内存在反光效果较好的物体(如镜子等)造成的镜面干扰;由于激光采集设备自身等因素造成的杂点干扰;由于人或物体位置的连续变化造成的动态干扰;由于待构建室内地图的空间中摆放、悬挂的小尺寸物体而造成的小尺寸物体干扰。无疑,这些干扰的存在都会影响所构 建的室内地图的精度。
图6为本发明实施例四提供的一种室内地图构建方法的流程图。本实施例在上述各实施例的基础上,进一步增加了滤波处理步骤。
本实施例提供的室内地图构建方法具体包括如下步骤:
S110、获取激光点云设备采集的激光点云数据。
S410、对所述激光点云数据包含的镜面干扰和/或杂点干扰进行滤波。
S420、在基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,同时对由于人或物体位置的连续变化造成的动态干扰进行滤波;依据配准结果和滤波结果,对激光点云数据进行拼合形成室内三维图。
S430、对所述室内三维图中的杂点干扰、小尺寸物体干扰或由于人或物体位置的连续变化造成的动态干扰进行滤波。
S130、对所述室内三维图进行切割形成室内地图。
需要说明的,在利于本实施例上述技术方案构建室内地图时,可以针对具体情况,有针对性地选择S410、S420以及S430之中的一个或多个步骤进行滤波处理。另外,由于干扰的类型不同,所需要的滤波算法可能相同,也可能不同。具体进行滤波时,根据干扰的类型选择相应的滤波算法进行滤波。例如,对于杂点干扰,可以选择离群点移除方法对激光点云数据进行滤波。
本实施例技术方案中,通过增加滤波处理步骤,可以有效解决由于干扰的存在导致所形成的室内地图精度低的问题,可以达到进一步提高所构建的室内地图精度的目的。
此外,还需要说明的是,在本申请中实施例一至实施例四从多个角度提供了多种用于提高所构建的室内地图的精度的方法。各方法在执行的过程中互不影响,可以互相叠加使用。
实施例五
图7为本发明实施例五提供的一种室内地图构建装置。该室内地图构建装置,包括:激光点云数据获取模块510、三维图构建模块520以及室内地图构建模块530。
激光点云数据获取模块510,用于获取激光点云设备采集的激光点云数据;
三维图构建模块520,用于基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图;
室内地图构建模块530,用于对所述室内三维图进行切割形成室内地图。
本实施例通过基于关键帧点云数据,对所获取的激光点云数据进行配准,形成室内三维图,并对所形成的室内三维图进行切割形成室内地图,可以解决目前在无建筑物的CAD图纸的情况下,构建室内地图的过程中,需要大量具备专业绘图能力的工作人员亲自对室内环境进行测绘,对工作人员能力要求高且劳动量大的问题,实现了降低对工作人员能力的要求,减小工作人员的劳动量,提高所构建的室内地图的精度的目的。
进一步地,该室内地图构建装置还可以包括:
激光点云修正模块,用于在基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准之前,获取惯性测量单元采集到的惯性测量单元姿态,其中所述惯性测量单元与所述激光点云设备刚性连接;依据所述惯性测量单元姿态,修正所述激光点云数据。
进一步地,该所述三维图构建模块520还可以包括:
配准单元,用于基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准;
移动轨迹确定单元,用于根据配准结果确定在采集所述激光点云数据的过程中所述激光点云设备的移动轨迹;
激光点云数据校正单元,用于确定所述移动轨迹是否构成闭合图形,若所述移动轨迹构成闭合图形,则利用回环优化算法,校正所述激光点云数据;
第一三维图形成单元,用于拼合校正后的激光点云数据,形成室内三维图。
进一步地,所述三维图构建模块520可以包括:
第一配准单元,用于基于所述激光点云数据包含的第一关键帧点云数据,对所述激光点云数据进行第一次配准;
第二配准单元,用于基于所述激光点云数据包含的第二关键帧点云数据,对第一次配准后的激光点云数据进行第二次配准,其中所述第二关键帧点云数 据与所述第一关键帧点云数据不同;
第二三维图形成单元,用于根据第二次配准结果,形成室内三维图。
进一步地,所述关键帧点云数据通过如下方式得到:
依据所述激光点云数据的采集场景,确定与该场景对应的激光点云数据的关键帧点云数据;或者,
对于每一帧激光点云数据,将前一帧激光点云数据确定为该帧激光点云数据的关键帧点云数据;或者,
依据所述激光点云数据的采集时间,确定所述激光点云数据包含的关键帧点云数据。
进一步地,室内地图构建装置还可以包括:
第一滤波模块,用于在基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准之前,对所述激光点云数据包含的镜面干扰和/或杂点干扰进行滤波。
进一步地,所述三维图构建模块520具体用于基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,且在配准过程中对由于人或物体位置的连续变化造成的动态干扰进行滤波;依据配准结果和滤波结果,对激光点云数据进行拼合形成室内三维图。
进一步地,该室内地图构建装置,还可以包括:
第二滤波模块,用于在对所述室内三维图进行切割形成室内地图之前,对所述室内三维图中的杂点干扰、小尺寸物体干扰或由于人或物体位置的连续变化造成的动态干扰进行滤波。
进一步地,所述室内地图构建模块530,包括:
室内空间划分单元,用于根据预设的分区规则,将室内空间划分为不同室内区域;
三维图切割单元,用于依据所述室内区域的区域特征,对所述室内三维图进行切割形成室内地图,所述区域特征包括高度。
上述产品可执行本发明任意实施例所提供的方法,具备执行方法相应的功能模块和有益效果。
实施例六
一种非易失性计算机存储介质,所述计算机存储介质存储有一个或者多个模块,当所述一个或者多个模块被一个执行室内地图构建方法的设备执行时,使得所述设备执行如下操作:
获取激光点云设备采集的激光点云数据;
基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图;
对所述室内三维图进行切割形成室内地图。
上述存储介质中存储的模块被所述设备执行时,基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准之前,还可以包括:
获取惯性测量单元采集到的惯性测量单元姿态,其中所述惯性测量单元与所述激光点云设备刚性连接;
依据所述惯性测量单元姿态,修正所述激光点云数据。
上述存储介质中存储的模块被所述设备执行时,基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图,可以包括:
基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准;
根据配准结果确定在采集所述激光点云数据的过程中所述激光点云设备的移动轨迹;
确定所述移动轨迹是否构成闭合图形,若所述移动轨迹构成闭合图形,则利用回环优化算法,校正所述激光点云数据;
拼合校正后的激光点云数据,形成室内三维图。
上述存储介质中存储的模块被所述设备执行时,基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图,可以包括:
基于所述激光点云数据包含的第一关键帧点云数据,对所述激光点云数据 进行第一次配准;
基于所述激光点云数据包含的第二关键帧点云数据,对第一次配准后的激光点云数据进行第二次配准,其中所述第二关键帧点云数据与所述第一关键帧点云数据不同;
根据第二次配准结果,形成室内三维图。
上述存储介质中存储的模块被所述设备执行时,关键帧点云数据可以通过如下方式得到:
依据所述激光点云数据的采集场景,确定与该场景对应的激光点云数据的关键帧点云数据;或者,
对于每一帧激光点云数据,将前一帧激光点云数据确定为该帧激光点云数据的关键帧点云数据;或者,
依据所述激光点云数据的采集时间,确定所述激光点云数据包含的关键帧点云数据。
上述存储介质中存储的模块被所述设备执行时,基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准之前,还可以包括:
对所述激光点云数据包含的镜面干扰和/或杂点干扰进行滤波。
上述存储介质中存储的模块被所述设备执行时,基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图,可以包括:
基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,且在配准过程中对由于人或物体位置的连续变化造成的动态干扰进行滤波;
依据配准结果和滤波结果,对激光点云数据进行拼合形成室内三维图。
上述存储介质中存储的模块被所述设备执行时,对所述室内三维图进行切割形成室内地图之前,还可以包括:
对所述室内三维图中的杂点干扰、小尺寸物体干扰或由于人或物体位置的连续变化造成的动态干扰进行滤波。
上述存储介质中存储的模块被所述设备执行时,对所述室内三维图进行切 割形成室内地图,可以包括:
根据预设的分区规则,将室内空间划分为不同室内区域;
依据所述室内区域的区域特征,对所述室内三维图进行切割形成室内地图,所述区域特征包括高度。
实施例七
请参阅图8,为本发明实施例七提供的一种执行室内地图构建方法的设备的硬件结构示意图。
该设备包括:
一个或者多个处理器610,图8中以一个处理器610为例;
存储器620;以及一个或者多个模块。
所述设备还可以包括:输入装置630和输出装置640。所述设备中的处理器610、存储器620、输入装置630和输出装置640可以通过总线或其他方式连接,图8中以通过总线连接为例。
存储器620作为一种计算机可读存储介质,可用于存储软件程序、计算机可执行程序以及模块,如本发明实施例中的室内地图构建方法对应的程序指令/模块(例如,附图7所示的室内地图构建装置中的激光点云数据获取模块510、三维图构建模块520以及室内地图构建模块530)。处理器610通过运行存储在存储器620中的软件程序、指令以及模块,从而执行服务器的各种功能应用以及数据处理,即实现上述方法实施例中的室内地图构建方法。
存储器620可包括存储程序区和存储数据区,其中,存储程序区可存储操作系统、至少一个功能所需的应用程序;存储数据区可存储根据终端设备的使用所创建的数据等。此外,存储器620可以包括高速随机存取存储器,还可以包括非易失性存储器,例如至少一个磁盘存储器件、闪存器件、或其他非易失性固态存储器件。在一些实例中,存储器620可进一步包括相对于处理器610远程设置的存储器,这些远程存储器可以通过网络连接至终端设备。上述网络的实例包括但不限于互联网、企业内部网、局域网、移动通信网及其组合。
输入装置630可用于接收输入的数字或字符信息,以及产生与终端的用户 设置以及功能控制有关的信号输入。输出装置640可包括显示屏等显示设备。
所述一个或者多个模块存储在所述存储器620中,当被所述一个或者多个处理器610执行时,执行如下操作:
获取激光点云设备采集的激光点云数据;
基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图;
对所述室内三维图进行切割形成室内地图。
进一步地,基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准之前,还包括:
获取惯性测量单元采集到的惯性测量单元姿态,其中所述惯性测量单元与所述激光点云设备刚性连接;
依据所述惯性测量单元姿态,修正所述激光点云数据。
进一步地,基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图,包括:
基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准;
根据配准结果确定在采集所述激光点云数据的过程中所述激光点云设备的移动轨迹;
确定所述移动轨迹是否构成闭合图形,若所述移动轨迹构成闭合图形,则利用回环优化算法,校正所述激光点云数据;
拼合校正后的激光点云数据,形成室内三维图。
进一步地,基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图,包括:
基于所述激光点云数据包含的第一关键帧点云数据,对所述激光点云数据进行第一次配准;
基于所述激光点云数据包含的第二关键帧点云数据,对第一次配准后的激光点云数据进行第二次配准,其中所述第二关键帧点云数据与所述第一关键帧点云数据不同;
根据第二次配准结果,形成室内三维图。
进一步地,关键帧点云数据通过如下方式得到:
依据所述激光点云数据的采集场景,确定与该场景对应的激光点云数据的关键帧点云数据;或者,
对于每一帧激光点云数据,将前一帧激光点云数据确定为该帧激光点云数据的关键帧点云数据;或者,
依据所述激光点云数据的采集时间,确定所述激光点云数据包含的关键帧点云数据。
进一步地,基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准之前,还包括:
对所述激光点云数据包含的镜面干扰和/或杂点干扰进行滤波。
进一步地,基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图,包括:
基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,且在配准过程中对由于人或物体位置的连续变化造成的动态干扰进行滤波;
依据配准结果和滤波结果,对激光点云数据进行拼合形成室内三维图。
进一步地,对所述室内三维图进行切割形成室内地图之前,还包括:
对所述室内三维图中的杂点干扰、小尺寸物体干扰或由于人或物体位置的连续变化造成的动态干扰进行滤波。
进一步地,对所述室内三维图进行切割形成室内地图,包括:
根据预设的分区规则,将室内空间划分为不同室内区域;
依据所述室内区域的区域特征,对所述室内三维图进行切割形成室内地图,所述区域特征包括高度。
通过以上关于实施方式的描述,所属领域的技术人员可以清楚地了解到,本发明可借助软件及必需的通用硬件来实现,当然也可以通过硬件实现,但很多情况下前者是更佳的实施方式。基于这样的理解,本发明的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来,该计算机 软件产品可以存储在计算机可读存储介质中,如计算机的软盘、只读存储器(Read-Only Memory,ROM)、随机存取存储器(Random Access Memory,RAM)、闪存(FLASH)、硬盘或光盘等,包括若干指令用以使得一台计算机设备(可以是个人计算机,服务器,或者网络设备等)执行本发明各个实施例所述的方法。
值得注意的是,上述室内地图构建装置的实施例中,所包括的各个单元和模块只是按照功能逻辑进行划分的,但并不局限于上述的划分,只要能够实现相应的功能即可;另外,各功能单元的具体名称也只是为了便于相互区分,并不用于限制本发明的保护范围。
以上所述,仅为本发明的具体实施方式,但本发明的保护范围并不局限于此,任何熟悉本技术领域的技术人员在本发明披露的技术范围内,可轻易想到的变化或替换,都应涵盖在本发明的保护范围之内。因此,本发明的保护范围应以所述权利要求的保护范围为准。

Claims (19)

  1. 一种室内地图构建方法,其特征在于,包括:
    获取激光点云设备采集的激光点云数据;
    基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图;
    对所述室内三维图进行切割形成室内地图。
  2. 根据权利要求1所述的方法,其特征在于,基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准之前,还包括:
    获取惯性测量单元采集到的惯性测量单元姿态,其中所述惯性测量单元与所述激光点云设备刚性连接;
    依据所述惯性测量单元姿态,修正所述激光点云数据。
  3. 根据权利要求1所述的方法,其特征在于,基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图,包括:
    基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准;
    根据配准结果确定在采集所述激光点云数据的过程中所述激光点云设备的移动轨迹;
    确定所述移动轨迹是否构成闭合图形,若所述移动轨迹构成闭合图形,则利用回环优化算法,校正所述激光点云数据;
    拼合校正后的激光点云数据,形成室内三维图。
  4. 根据权利要求1所述的方法,其特征在于,基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图,包括:
    基于所述激光点云数据包含的第一关键帧点云数据,对所述激光点云数据进行第一次配准;
    基于所述激光点云数据包含的第二关键帧点云数据,对第一次配准后的激光点云数据进行第二次配准,其中所述第二关键帧点云数据与所述第一关键帧点云数据不同;
    根据第二次配准结果,形成室内三维图。
  5. 根据权利要求1所述的方法,其特征在于,关键帧点云数据通过如下方式得到:
    依据所述激光点云数据的采集场景,确定与该场景对应的激光点云数据的关键帧点云数据;或者,
    对于每一帧激光点云数据,将前一帧激光点云数据确定为该帧激光点云数据的关键帧点云数据;或者,
    依据所述激光点云数据的采集时间,确定所述激光点云数据包含的关键帧点云数据。
  6. 根据权利要求1所述的方法,其特征在于,基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准之前,还包括:
    对所述激光点云数据包含的镜面干扰和/或杂点干扰进行滤波。
  7. 根据权利要求1所述的方法,其特征在于,基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图,包括:
    基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,且在配准过程中对由于人或物体位置的连续变化造成的动态干扰进行滤波;
    依据配准结果和滤波结果,对激光点云数据进行拼合形成室内三维图。
  8. 根据权利要求1所述的方法,其特征在于,对所述室内三维图进行切割形成室内地图之前,还包括:
    对所述室内三维图中的杂点干扰、小尺寸物体干扰或由于人或物体位置的连续变化造成的动态干扰进行滤波。
  9. 根据权利要求1所述的方法,其特征在于,对所述室内三维图进行切割形成室内地图,包括:
    根据预设的分区规则,将室内空间划分为不同室内区域;
    依据所述室内区域的区域特征,对所述室内三维图进行切割形成室内地图,所述区域特征包括高度。
  10. 一种室内地图构建装置,其特征在于,包括:
    激光点云数据获取模块,用于获取激光点云设备采集的激光点云数据;
    三维图构建模块,用于基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图;
    室内地图构建模块,用于对所述室内三维图进行切割形成室内地图。
  11. 根据权利要求10所述的装置,其特征在于,还包括:
    激光点云修正模块,用于在基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准之前,获取惯性测量单元采集到的惯性测量单元姿态,其中所述惯性测量单元与所述激光点云设备刚性连接;依据所述惯性测量单元姿态,修正所述激光点云数据。
  12. 根据权利要求10所述的装置,其特征在于,所述三维图构建模块包括:
    配准单元,用于基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准;
    移动轨迹确定单元,用于根据配准结果确定在采集所述激光点云数据的过程中所述激光点云设备的移动轨迹;
    激光点云数据校正单元,用于确定所述移动轨迹是否构成闭合图形,若所述移动轨迹构成闭合图形,则利用回环优化算法,校正所述激光点云数据;
    第一三维图形成单元,用于拼合校正后的激光点云数据,形成室内三维图。
  13. 根据权利要求10所述的装置,其特征在于,所述三维图构建模块包括:
    第一配准单元,用于基于所述激光点云数据包含的第一关键帧点云数据,对所述激光点云数据进行第一次配准;
    第二配准单元,用于基于所述激光点云数据包含的第二关键帧点云数据,对第一次配准后的激光点云数据进行第二次配准,其中所述第二关键帧点云数据与所述第一关键帧点云数据不同;
    第二三维图形成单元,用于根据第二次配准结果,形成室内三维图。
  14. 根据权利要求10所述的装置,其特征在于,所述关键帧点云数据通 过如下方式得到:
    依据所述激光点云数据的采集场景,确定与该场景对应的激光点云数据的关键帧点云数据;或者,
    对于每一帧激光点云数据,将前一帧激光点云数据确定为该帧激光点云数据的关键帧点云数据;或者,
    依据所述激光点云数据的采集时间,确定所述激光点云数据包含的关键帧点云数据。
  15. 根据权利要求10所述的装置,其特征在于,还包括:
    第一滤波模块,用于在基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准之前,对所述激光点云数据包含的镜面干扰和/或杂点干扰进行滤波。
  16. 根据权利要求10所述的装置,其特征在于,所述三维图构建模块具体用于基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,且在配准过程中对由于人或物体位置的连续变化造成的动态干扰进行滤波;依据配准结果和滤波结果,对激光点云数据进行拼合形成室内三维图。
  17. 根据权利要求10所述的装置,其特征在于,还包括:
    第二滤波模块,用于在对所述室内三维图进行切割形成室内地图之前,对所述室内三维图中的杂点干扰、小尺寸物体干扰或由于人或物体位置的连续变化造成的动态干扰进行滤波。
  18. 根据权利要求10所述的装置,其特征在于,所述室内地图构建模块,包括:
    室内空间划分单元,用于根据预设的分区规则,将室内空间划分为不同室内区域;
    三维图切割单元,用于依据所述室内区域的区域特征,对所述室内三维图进行切割形成室内地图,所述区域特征包括高度。
  19. 一种包含计算机可执行指令的存储介质,所述计算机可执行指令在由计算机处理器执行时用于执行一种室内地图构建方法,其特征在于,该方法包括:
    获取激光点云设备采集的激光点云数据;
    基于所述激光点云数据包含的关键帧点云数据,对所述激光点云数据进行配准,并依据配准结果形成室内三维图;
    对所述室内三维图进行切割形成室内地图。
PCT/CN2016/096104 2016-03-31 2016-08-19 一种室内地图构建方法、装置和存储介质 WO2017166594A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201610202360.9 2016-03-31
CN201610202360.9A CN105702151B (zh) 2016-03-31 2016-03-31 一种室内地图构建方法及装置

Publications (1)

Publication Number Publication Date
WO2017166594A1 true WO2017166594A1 (zh) 2017-10-05

Family

ID=56219090

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/096104 WO2017166594A1 (zh) 2016-03-31 2016-08-19 一种室内地图构建方法、装置和存储介质

Country Status (2)

Country Link
CN (1) CN105702151B (zh)
WO (1) WO2017166594A1 (zh)

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109285220A (zh) * 2018-08-30 2019-01-29 百度在线网络技术(北京)有限公司 一种三维场景地图的生成方法、装置、设备及存储介质
CN110363847A (zh) * 2018-04-10 2019-10-22 北京京东尚科信息技术有限公司 一种基于点云数据的地图模型构建方法和装置
CN110415342A (zh) * 2019-08-02 2019-11-05 深圳市唯特视科技有限公司 一种基于多融合传感器的三维点云重建装置与方法
CN111009005A (zh) * 2019-11-27 2020-04-14 天津大学 几何信息与光度信息相结合的场景分类点云粗配准方法
CN111476830A (zh) * 2020-03-13 2020-07-31 上海高仙自动化科技发展有限公司 点云数据处理方法、机器人、电子设备和可读存储介质
CN111767767A (zh) * 2019-05-24 2020-10-13 北京京东尚科信息技术有限公司 室内物体的检测方法、装置、电子设备与存储介质
CN111915662A (zh) * 2019-05-07 2020-11-10 北京京东尚科信息技术有限公司 一种三维激光点云数据预处理方法和装置
CN112540616A (zh) * 2020-12-11 2021-03-23 北京赛目科技有限公司 一种基于无人驾驶的激光点云生成方法及装置
CN112818898A (zh) * 2021-02-20 2021-05-18 北京字跳网络技术有限公司 模型训练方法、装置和电子设备
CN113379910A (zh) * 2021-06-09 2021-09-10 山东大学 基于slam的移动机器人矿山场景重建方法及系统
CN113739785A (zh) * 2020-05-29 2021-12-03 杭州海康机器人技术有限公司 一种机器人定位方法、装置及存储介质
CN114998522A (zh) * 2022-06-15 2022-09-02 中国测绘科学研究院 多视连续光场影像室内场景稠密点云准确提取方法及系统
CN115223442A (zh) * 2021-07-22 2022-10-21 上海数川数据科技有限公司 一种室内行人地图的自动生成方法
WO2024089140A1 (fr) * 2022-10-26 2024-05-02 Offroad Système et méthode de construction d'une carte d'un environnement

Families Citing this family (27)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105702151B (zh) * 2016-03-31 2019-06-11 百度在线网络技术(北京)有限公司 一种室内地图构建方法及装置
CN107631688A (zh) * 2016-07-18 2018-01-26 巧夺天宫(深圳)科技有限公司 基于三维激光点云的装修测量方法和系统
CN106504275B (zh) * 2016-10-12 2019-03-05 杭州深瞳科技有限公司 一种惯性定位与点云配准耦合互补的实时三维重建方法
CN106846308B (zh) * 2017-01-20 2019-08-20 广州市城市规划勘测设计研究院 基于点云的地形图精度的检测方法和装置
CN106969763B (zh) * 2017-04-07 2021-01-01 百度在线网络技术(北京)有限公司 用于确定无人驾驶车辆的偏航角的方法和装置
CN108734780B (zh) * 2017-04-17 2021-12-10 百度在线网络技术(北京)有限公司 用于生成地图的方法、装置和设备
CN113341397A (zh) * 2017-08-15 2021-09-03 百度在线网络技术(北京)有限公司 反射值地图构建方法和装置
CN107843208A (zh) * 2017-10-27 2018-03-27 北京矿冶研究总院 一种矿山巷道轮廓感知方法及系统
CN108230247B (zh) * 2017-12-29 2019-03-15 达闼科技(北京)有限公司 基于云端的三维地图的生成方法、装置、设备及计算机可读的存储介质
US20200081444A1 (en) * 2018-01-12 2020-03-12 Hong Huang Cleaning method capable of automatically planning path
WO2019136714A1 (zh) * 2018-01-12 2019-07-18 浙江国自机器人技术有限公司 一种基于3d激光的地图构建方法及系统
CN108550318B (zh) * 2018-03-12 2020-09-29 浙江大华技术股份有限公司 一种构建地图的方法及装置
CN110363849A (zh) * 2018-04-11 2019-10-22 株式会社日立制作所 一种室内三维建模方法及系统
CN110634183A (zh) * 2018-06-21 2019-12-31 北京京东尚科信息技术有限公司 地图构建方法、装置和无人设备
CN110376605B (zh) * 2018-09-18 2022-02-01 北京京东尚科信息技术有限公司 地图构建方法、导航方法和装置
CN109297496A (zh) * 2018-09-29 2019-02-01 上海新世纪机器人有限公司 基于slam的机器人定位方法及装置
CN109949414A (zh) * 2019-01-31 2019-06-28 顺丰科技有限公司 室内地图的构建方法及装置
CN110120093A (zh) * 2019-03-25 2019-08-13 深圳大学 一种多元特征混合优化的rgb-d室内三维测图方法及系统
CN110084840B (zh) 2019-04-24 2022-05-13 阿波罗智能技术(北京)有限公司 点云配准方法、装置、服务器和计算机可读介质
CN109934920B (zh) * 2019-05-20 2019-08-09 奥特酷智能科技(南京)有限公司 基于低成本设备的高精度三维点云地图构建方法
CN111239763A (zh) * 2020-03-06 2020-06-05 广州视源电子科技股份有限公司 对象的定位方法、装置、存储介质和处理器
CN111412842B (zh) * 2020-04-09 2022-02-25 广东博智林机器人有限公司 墙面的截面尺寸的测量方法及装置、系统
CN113551678A (zh) * 2020-04-09 2021-10-26 阿里巴巴集团控股有限公司 构建地图的方法、构建高精度地图的方法及移动设备
CN111553937B (zh) * 2020-04-23 2023-11-21 东软睿驰汽车技术(上海)有限公司 激光点云地图构建方法、装置、设备及系统
CN111928839A (zh) * 2020-08-07 2020-11-13 北京星天地信息科技有限公司 一种通行路线规划方法、装置及计算机设备
CN113693493B (zh) * 2021-02-10 2023-03-10 北京石头创新科技有限公司 区域地图绘制方法及装置、介质及电子设备
CN115439644B (zh) * 2022-08-19 2023-08-08 广东领慧数字空间科技有限公司 一种相似点云数据对齐方法

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103148804A (zh) * 2013-03-04 2013-06-12 清华大学 一种基于激光扫描的室内未知结构识别方法
CN103278170A (zh) * 2013-05-16 2013-09-04 东南大学 基于显著场景点检测的移动机器人级联地图创建方法
CN103941264A (zh) * 2014-03-26 2014-07-23 南京航空航天大学 一种室内未知环境下使用激光雷达定位方法
CN103941750A (zh) * 2014-04-30 2014-07-23 东北大学 基于小型四旋翼无人机的构图装置及方法
CN104897161A (zh) * 2015-06-02 2015-09-09 武汉大学 基于激光测距的室内平面地图制图方法
CN104964683A (zh) * 2015-06-04 2015-10-07 上海物景智能科技有限公司 一种室内环境地图创建的闭环校正方法
CN105222789A (zh) * 2015-10-23 2016-01-06 哈尔滨工业大学 一种基于激光测距传感器的楼宇室内平面图建立方法
CN105354875A (zh) * 2015-09-25 2016-02-24 厦门大学 一种室内环境二维与三维联合模型的构建方法和系统
CN105702151A (zh) * 2016-03-31 2016-06-22 百度在线网络技术(北京)有限公司 一种室内地图构建方法及装置

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100757751B1 (ko) * 2006-06-01 2007-09-11 한국과학기술연구원 실내 환경의 환경맵 생성 장치 및 방법
CN104778688B (zh) * 2015-03-27 2018-03-13 华为技术有限公司 点云数据的配准方法及装置

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103148804A (zh) * 2013-03-04 2013-06-12 清华大学 一种基于激光扫描的室内未知结构识别方法
CN103278170A (zh) * 2013-05-16 2013-09-04 东南大学 基于显著场景点检测的移动机器人级联地图创建方法
CN103941264A (zh) * 2014-03-26 2014-07-23 南京航空航天大学 一种室内未知环境下使用激光雷达定位方法
CN103941750A (zh) * 2014-04-30 2014-07-23 东北大学 基于小型四旋翼无人机的构图装置及方法
CN104897161A (zh) * 2015-06-02 2015-09-09 武汉大学 基于激光测距的室内平面地图制图方法
CN104964683A (zh) * 2015-06-04 2015-10-07 上海物景智能科技有限公司 一种室内环境地图创建的闭环校正方法
CN105354875A (zh) * 2015-09-25 2016-02-24 厦门大学 一种室内环境二维与三维联合模型的构建方法和系统
CN105222789A (zh) * 2015-10-23 2016-01-06 哈尔滨工业大学 一种基于激光测距传感器的楼宇室内平面图建立方法
CN105702151A (zh) * 2016-03-31 2016-06-22 百度在线网络技术(北京)有限公司 一种室内地图构建方法及装置

Cited By (22)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110363847A (zh) * 2018-04-10 2019-10-22 北京京东尚科信息技术有限公司 一种基于点云数据的地图模型构建方法和装置
CN110363847B (zh) * 2018-04-10 2023-06-23 北京京东尚科信息技术有限公司 一种基于点云数据的地图模型构建方法和装置
CN109285220A (zh) * 2018-08-30 2019-01-29 百度在线网络技术(北京)有限公司 一种三维场景地图的生成方法、装置、设备及存储介质
CN111915662B (zh) * 2019-05-07 2024-01-12 北京京东乾石科技有限公司 一种三维激光点云数据预处理方法和装置
CN111915662A (zh) * 2019-05-07 2020-11-10 北京京东尚科信息技术有限公司 一种三维激光点云数据预处理方法和装置
CN111767767A (zh) * 2019-05-24 2020-10-13 北京京东尚科信息技术有限公司 室内物体的检测方法、装置、电子设备与存储介质
CN110415342B (zh) * 2019-08-02 2023-04-18 深圳市唯特视科技有限公司 一种基于多融合传感器的三维点云重建装置与方法
CN110415342A (zh) * 2019-08-02 2019-11-05 深圳市唯特视科技有限公司 一种基于多融合传感器的三维点云重建装置与方法
CN111009005A (zh) * 2019-11-27 2020-04-14 天津大学 几何信息与光度信息相结合的场景分类点云粗配准方法
CN111476830A (zh) * 2020-03-13 2020-07-31 上海高仙自动化科技发展有限公司 点云数据处理方法、机器人、电子设备和可读存储介质
CN111476830B (zh) * 2020-03-13 2023-08-01 上海高仙自动化科技发展有限公司 点云数据处理方法、机器人、电子设备和可读存储介质
CN113739785A (zh) * 2020-05-29 2021-12-03 杭州海康机器人技术有限公司 一种机器人定位方法、装置及存储介质
CN112540616A (zh) * 2020-12-11 2021-03-23 北京赛目科技有限公司 一种基于无人驾驶的激光点云生成方法及装置
CN112540616B (zh) * 2020-12-11 2021-07-16 北京赛目科技有限公司 一种基于无人驾驶的激光点云生成方法及装置
CN112818898B (zh) * 2021-02-20 2024-02-20 北京字跳网络技术有限公司 模型训练方法、装置和电子设备
CN112818898A (zh) * 2021-02-20 2021-05-18 北京字跳网络技术有限公司 模型训练方法、装置和电子设备
CN113379910A (zh) * 2021-06-09 2021-09-10 山东大学 基于slam的移动机器人矿山场景重建方法及系统
CN115223442A (zh) * 2021-07-22 2022-10-21 上海数川数据科技有限公司 一种室内行人地图的自动生成方法
CN115223442B (zh) * 2021-07-22 2024-04-09 上海数川数据科技有限公司 一种室内行人地图的自动生成方法
CN114998522A (zh) * 2022-06-15 2022-09-02 中国测绘科学研究院 多视连续光场影像室内场景稠密点云准确提取方法及系统
WO2024089140A1 (fr) * 2022-10-26 2024-05-02 Offroad Système et méthode de construction d'une carte d'un environnement
FR3141535A1 (fr) * 2022-10-26 2024-05-03 Offroad système et méthode de construction d’une carte d’un environnement

Also Published As

Publication number Publication date
CN105702151A (zh) 2016-06-22
CN105702151B (zh) 2019-06-11

Similar Documents

Publication Publication Date Title
WO2017166594A1 (zh) 一种室内地图构建方法、装置和存储介质
US10809066B2 (en) Automated mapping information generation from inter-connected images
Dong et al. ViNav: A vision-based indoor navigation system for smartphones
JP2021152662A (ja) リアルタイムのマッピングと位置確認のための方法及び装置
CN110455291B (zh) 用于利用磁场异常进行户内定位-寻找的方法和装置
CN110118560B (zh) 一种基于lstm和多传感器融合的室内定位方法
CN109461208B (zh) 三维地图处理方法、装置、介质和计算设备
CN109084732A (zh) 定位与导航方法、装置及处理设备
Puyol et al. Pedestrian simultaneous localization and mapping in multistory buildings using inertial sensors
CN111402339B (zh) 一种实时定位方法、装置、系统及存储介质
CN105516929B (zh) 建立室内地图数据的方法、室内定位方法及对应装置
CN102622762A (zh) 使用深度图的实时相机跟踪
CN111595334B (zh) 基于视觉点线特征与imu紧耦合的室内自主定位方法
WO2011046971A1 (en) Method and apparatus for identification of points of interest within a predefined area
CN106997614A (zh) 一种基于深度相机的大规模场景3d建模方法及其装置
CN111862215B (zh) 一种计算机设备定位方法、装置、计算机设备和存储介质
CN113985429A (zh) 基于三维激光雷达的无人机环境扫描与重构方法
CN112950790A (zh) 路线导航方法、装置、电子设备及存储介质
WO2015079260A1 (en) Location finding apparatus and associated methods
US20240013484A1 (en) Method for generating roof outlines from lateral images
CN112105892A (zh) 使用运动数据和面元数据识别地图特征
CA3091599C (en) Method and system for lightweight calibration of magnetic fingerprint dataset
CN113610702B (zh) 一种建图方法、装置、电子设备及存储介质
Nyqvist et al. A high-performance tracking system based on camera and IMU
Luo et al. Indoor mapping using low-cost MLS point clouds and architectural skeleton constraints

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16896366

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 16896366

Country of ref document: EP

Kind code of ref document: A1