WO2020057462A1 - 列车内定位、以及室内定位 - Google Patents

列车内定位、以及室内定位 Download PDF

Info

Publication number
WO2020057462A1
WO2020057462A1 PCT/CN2019/105969 CN2019105969W WO2020057462A1 WO 2020057462 A1 WO2020057462 A1 WO 2020057462A1 CN 2019105969 W CN2019105969 W CN 2019105969W WO 2020057462 A1 WO2020057462 A1 WO 2020057462A1
Authority
WO
WIPO (PCT)
Prior art keywords
feature
environment image
frame
features
image
Prior art date
Application number
PCT/CN2019/105969
Other languages
English (en)
French (fr)
Inventor
聂琼
Original Assignee
北京三快在线科技有限公司
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 北京三快在线科技有限公司 filed Critical 北京三快在线科技有限公司
Priority to US17/276,823 priority Critical patent/US20210350142A1/en
Publication of WO2020057462A1 publication Critical patent/WO2020057462A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02—Services making use of location information
    • H04W4/021—Services related to particular areas, e.g. point of interest [POI] services, venue services or geofences
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02—Services making use of location information
    • H04W4/023—Services making use of location information using mutual or relative location information between multiple location based services [LBS] targets or of distance thresholds
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
    • G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
    • G01S5/0252—Radio frequency fingerprinting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/20—Analysis of motion
    • G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/50—Depth or shape recovery
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02—Services making use of location information
    • H04W4/029—Location-based management or tracking services
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/30—Services specially adapted for particular environments, situations or purposes
    • H04W4/40—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P]
    • H04W4/42—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P] for mass transport vehicles, e.g. buses, trains or aircraft
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30232—Surveillance
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30244—Camera pose
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30248—Vehicle exterior or interior
    • G06T2207/30268—Vehicle interior

Definitions

  • the present application relates to the field of positioning technology, and in particular, to positioning in trains and indoor positioning solutions.
  • SLAM simultaneous localization and mapping
  • a method for positioning in a train including:
  • a position of the mobile device in the train is determined based on the number of the features passed by the mobile device.
  • a positioning device in a train including:
  • An image acquisition unit for acquiring an environment image in a train
  • a feature determining unit for determining a feature based on the environment image, wherein the feature is set in the train according to a preset rule
  • a position determining unit is configured to determine a position of the mobile device in the train based on the number of the features passed by the mobile device.
  • a mobile device including:
  • Memory for storing processor-executable instructions
  • the processor is configured to execute the above-mentioned in-train positioning method.
  • a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the foregoing in-train positioning method is performed.
  • an indoor positioning method including: acquiring an indoor environment image; determining a feature based on the environment image, wherein the feature is set in the room according to a preset rule; A position of the mobile device in the room is determined based on the number of the features passed by the mobile device.
  • an indoor positioning device including: an image acquisition unit for acquiring an indoor environment image; and a feature determination unit for determining a feature based on the environment image, wherein The feature is set in the room according to a preset rule; a position determination unit determines a position of the mobile device in the room based on the number of the features passed by the mobile device.
  • a mobile device including: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to execute the indoor positioning method described above.
  • a computer-readable storage medium in which a computer program is stored, and the program is executed by a processor to execute the above-mentioned indoor positioning method.
  • the position of the robot in the room can be determined according to the number of passing features, and the features are extracted from the environmental image.
  • the features are extracted from the environmental image.
  • there is no need to build a map through SLAM so there is no need to set a lidar on the robot. Therefore, costs can be saved; on the other hand, only the number of features to be passed needs to be determined, and the coordinates of the robot in the map are relatively determined.
  • the calculation process is simple and time-consuming, thereby ensuring real-time positioning.
  • Fig. 1A is a schematic flowchart of an indoor positioning method according to an embodiment of the present application.
  • Fig. 1B is a schematic flowchart of a method for positioning in a train according to an embodiment of the present application.
  • Fig. 2 is a schematic flowchart of determining a position of a mobile device in the train based on the number of the features passed by the mobile device according to an embodiment of the present application.
  • Fig. 3 is a schematic flowchart of another in-train positioning method according to an embodiment of the present application.
  • Fig. 4 is a schematic flowchart of determining the number of passing features according to an embodiment of the present application.
  • Fig. 5 is a schematic flowchart of tracking a feature according to a preset manner according to an embodiment of the present application.
  • Fig. 6 is a schematic flowchart of updating a position of a feature in an environment image according to an embodiment of the present application.
  • Fig. 7 is a schematic flowchart illustrating whether tracking of a feature is ended according to an embodiment of the present application.
  • Fig. 8 is another schematic flowchart for determining whether tracking of a feature is ended according to an embodiment of the present application.
  • Fig. 9 is a hardware structural diagram of a robot in which a positioning device in a train is located according to an embodiment of the present application.
  • Fig. 10 is a schematic block diagram of a positioning device in a train according to an embodiment of the present application.
  • Fig. 11 is a schematic block diagram of a position determining unit according to an embodiment of the present application.
  • Fig. 12 is a schematic block diagram of another in-train positioning device according to an embodiment of the present application.
  • Fig. 13 is a schematic block diagram of another location determining unit according to an embodiment of the present application.
  • Fig. 14 is a schematic block diagram of a tracking subunit according to an embodiment of the present application.
  • Fig. 15 is a schematic block diagram of a location update module according to an embodiment of the present application.
  • first, second, third, etc. may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other.
  • first information may also be referred to as second information, and similarly, the second information may also be referred to as first information.
  • word “if” as used herein can be interpreted as “at” or "when” or "in response to determination”.
  • Fig. 1A is a schematic flowchart of an indoor positioning method according to an embodiment of the present application.
  • the method shown in this embodiment can be applied to a mobile device, such as a robot, which can move in a specific area within a train, for example.
  • Step S1 ' acquiring an indoor environment image
  • Step S2 ' determining a feature based on the environment image, wherein the feature is set in the room according to a preset rule
  • Step S3 ' determining a position of the mobile device in the room based on the number of the features passed by the mobile device.
  • the robot can perform tasks such as distribution (e.g., equipped with a loading device on the robot), ticket checking (e.g., equipped with a scanning device on the robot), cleaning (e.g., equipped with a cleaning device on the robot), and the like.
  • distribution e.g., equipped with a loading device on the robot
  • ticket checking e.g., equipped with a scanning device on the robot
  • cleaning e.g., equipped with a cleaning device on the robot
  • the in-train positioning method may include the following steps S1 to S3.
  • step S1 an environment image in the train is acquired.
  • an image acquisition device such as a depth camera
  • the captured environment image may be a depth image.
  • the train refers to a vehicle having a plurality of cars, such as a train, a subway, a high-speed rail, a train, and the like.
  • Step S2 Determine a feature based on the environment image.
  • the feature is set in the train according to a preset rule.
  • the features are set according to a preset rule in the train, and may include one or a combination of the following rules: the number of the features is positively related to the number of cars of the train The features are arranged in the train in a preset order, and the features are arranged at equal intervals in the train.
  • the bounding box of the image object can be obtained.
  • the bounding box can be understood as a rectangular box that surrounds the object in the image.
  • it can be based on the MASK-RCNN
  • the network performs an object detection algorithm) or SSD (an object detection algorithm that directly predicts the coordinates and categories of the bounding box) to process the obtained bounding box to determine whether it is a feature.
  • a recognition model can also be trained through machine learning, and features can be identified in the environment image according to the recognition model. The process of specifically determining the features is not the main improvement point of this embodiment, so it will not be repeated here.
  • Step S3 Determine the position of the mobile device in the train based on the number of the features passed by the mobile device.
  • the feature may be a car door, a window, a seat (which may be an entire seat, or a part of the structure of the seat, such as a handle, a seat back, etc.), a sleeper, or the like.
  • a car door each car is provided with a car door at each end.
  • Setting according to a preset rule may mean that the ratio of the number of car doors to the number of cars is 2 and the car door is along the train. The length of the carriage is set.
  • K is a positive integer
  • Setting according to a preset rule can mean that the ratio of the number of seats to the number of compartments is K and each row The product of the number of seats, and the seats are arranged in the train along the length of the carriage.
  • the features are set according to the preset rules in the train, when the robot moves in the train in a manner related to the preset rules, the more features the robot passes, the more distance it travels. Therefore, based on the distance traveled and preset rules, the position of the robot inside the train can be determined.
  • setting according to a preset rule means that the number of features is positively related to the number of cars, and features are arranged along the length of the cars in the train. Then, when the robot moves along the length of the carriage, the distance it travels is positively related to the number of features it passes, so the position in the train can be determined according to the number of features that the robot passed.
  • the starting position of the robot is at the front and moves towards the rear.
  • the compartment from the front to the rear is compartments 1 to 8.
  • Each compartment contains 20 rows of seats from the compartment head to the compartment.
  • Rear seats are installed from the first to the 20th row. If the number of features passed is 65, it can be determined that 3 cars have passed, are currently located in car 4 and 5 cars have passed through 5 rows of seats, and are currently located in 6 rows of seats.
  • the robot's position in the train is the sixth row of seats in carriage 4.
  • the position of the robot in the train can be determined according to the number of passing features, and the features are extracted from the environmental image.
  • the features are extracted from the environmental image.
  • the concept of this embodiment can also be applied to other space scenarios, such as movie theaters, workshops, warehouses, and so on.
  • the feature when applied to a movie theater, the feature may be a seat, and the position of the mobile device in the movie theater may be determined according to the number of passing seats.
  • the position of the mobile device in the workshop when used in the workshop, the position of the mobile device in the workshop can be determined according to the number of passing machine tools.
  • the position of the mobile device in the warehouse can be determined according to the number of stored items (such as boxes and barrels).
  • Fig. 2 is a schematic flowchart of determining a position of a mobile device in the train based on the number of the features passed by the mobile device according to an embodiment of the present application.
  • the features include a first feature and a second feature
  • the mobile device is determined based on the number of the features passed by the mobile device.
  • the position in the train includes:
  • Step S31 Determine a carriage in which the mobile device is located in the train according to the number of first features passed by the mobile device;
  • Step S32 Determine the position of the mobile device in the compartment in which it is located according to the number of second features that the mobile device passes in the compartment in which it is located.
  • the first feature may be a feature used to characterize a cabin.
  • the ratio of the number of the first features to the number of the compartments is less than or equal to 2, that is, the first feature will hardly appear repeatedly in the compartment.
  • the first feature is a compartment door, a toilet, or the like.
  • the ratio of the number of the second features to the number of the carriages is greater than 2, that is, the second feature will repeatedly appear in the carriage.
  • the second feature is a seat, a window, or the like.
  • the compartment in which the robot is located can be determined according to the number of passing second features, on the one hand, the compartment in which the robot is located is determined based on the amount of data of the second feature, and the calculation amount is large. For example, based on the embodiment shown in FIG. 1, the position that the robot needs to reach is the sixth row of seats in No. 4 compartment. In order to reach this position, the number of recorded passing seats is 65.
  • the compartment in which the robot is located may be determined according to the number of first features passing by the robot, and then the robot corresponding to the compartment in which it is located may be determined according to the number of second features through which the robot passes.
  • the position that the robot needs to reach is the 6th row of seats in No. 4 compartment
  • the number of passing compartment doors can be recorded first, that is, only 6 passing compartment doors (each compartment includes 2 compartment doors) can be recorded to determine It is located in No. 4 compartment, and then records the number of seats that the robot passed in No. 4 compartment. You only need to record after passing 5 rows of seats to determine the position of the robot in No. 4 compartment corresponding to the sixth row of seats.
  • Fig. 3 is a schematic flowchart of another in-train positioning method according to an embodiment of the present application. As shown in FIG. 3, based on the embodiment shown in FIG. 2, the method further includes:
  • Step S4 Determine the relative position of the mobile device and the second feature according to the distance between the mobile device and the second feature.
  • the position of the robot in the train can only be determined according to the number of passing features, there is a large error in this position. For example, taking the feature as the seat as an example, the distance between the two rows of seats is 1 meter, so when the robot passes the nth row of seats and has not passed the n + 1th row of seats, it can only determine that the position is in the first row. The area between the n-th row of seats and the n + 1th row of seats, so the robot determines a position error of about 1 meter.
  • the relative position of the mobile device and the second feature can be determined.
  • the obtained environment image is a depth image.
  • the distance from the seat to the robot can be determined, that is, the relative distance between the robot and the seat along the train length direction.
  • the error of the relative distance between the robot and the seat is much smaller than the distance between the two rows of seats. For example, when the robot passes the nth row of seats and has not passed the n + 1th row of seats, it can determine the relative distance between itself and the n + 1th row of seats.
  • the relative distance is 0.4 meters, then it can be determined that the robot is currently located between the nth and n + 1th rows of seats, and 0.4 meters to the n + 1th row of seats. This is more accurate than just determining that the area is between the nth row of seats and the n + 1th row of seats, which facilitates more accurate positioning of the robot.
  • Fig. 4 is a schematic flowchart of determining the number of passing features according to an embodiment of the present application. As shown in FIG. 4, on the basis of the embodiment shown in FIG. 2, the number of the passed features is determined by the following methods:
  • Step S33 Track the feature according to a preset method
  • Step S34 it is determined whether the tracking of the feature is ended, and if it is finished, it is determined to pass the feature;
  • step S35 the number of passing features is updated.
  • the feature may be tracked in a preset manner.
  • the preset methods can be selected according to requirements. In the following embodiments, two preset methods are mainly described by way of example.
  • one preset method is to end the tracking of a feature when the feature meets a specific condition in the environment image
  • another preset method is when the feature does not exist in the environment image (for example, from the environment of the current frame). When the image disappears), the tracking of the feature is ended.
  • a feature can be determined and the number of features can be updated. For example, each time a feature is passed, 1 is added to the number of currently recorded features. By analogy, the number of passing features can be determined, and the position of the robot can be determined according to the number.
  • Fig. 5 is a schematic flowchart of tracking a feature according to a preset manner according to an embodiment of the present application. As shown in FIG. 5, on the basis of the embodiment shown in FIG. 4, tracking the feature according to a preset manner includes:
  • Step S331 Determine whether the feature in the n-th frame of the environment image is the same as the feature in the n + 1th frame of the environment image, and n is a positive integer;
  • Step S332 if the features are the same, update the positions of the features in the environment image according to the n + 1 frame environment image;
  • step S333 if the features are not the same, the features in the n + 1 frame environment image are tracked according to the preset method.
  • multiple frames of environmental images can be acquired continuously.
  • the features in the n + 1 frame of the environment image can be determined, and the features in the n + 1 frame of the environment image are compared with the features in the n frame of the environment image Compare, for example, compare the bounding boxes of features. Among them, NCC (Normalization Cross Correlation) can be used for comparison. In the comparison process, the position of the feature (such as the position of the center of the bounding box) and the motion speed of the robot can also be considered.
  • NCC Normalization Cross Correlation
  • the acquisition time of the environmental image of the nth frame and the n + 1th frame is separated by 0.1 second
  • the robot's movement speed is 0.5 m / s
  • the features in the nth frame of the environmental image and the features in the n + 1th frame of the environmental image are separated.
  • the position relative to the robot differs by 1 meter, which is far greater than 0.05 meters.
  • it can be determined that the features in the n + 1 frame environment image and the features in the n frame environment image are different features. In other words, if the feature in the n + 1 frame environment image and the feature in the n frame environment image are the same feature, the position of the two features relative to the robot differs by about 0.05 meters.
  • the comparison results determine that the feature in the n + 1 frame environment image and the feature in the n frame environment image are the same feature, that is, the feature has appeared in the n frame environment image, and in the n + 1 frame It also appears in the environment image, so the position of the feature in the environment image can be updated according to the position of the feature in the n + 1 frame environment image. In this way, the position of the stored feature can be guaranteed to correspond to the recently acquired environmental image, so that the area where the feature is located in the environmental image can be accurately determined for each subsequent frame of environmental image, and then where the mobile device is located A frame of environmental images passes through the feature.
  • the feature in the n + 1 frame environment image and the feature in the n frame environment image are not the same feature, that is, the feature does not appear in the n frame environment image, but in the nth frame Appears in the +1 frame environmental image, indicating that the feature is a new feature in the n + 1 frame environmental image, so the newly appeared feature can be tracked, tracking method and tracking method of the aforementioned feature The same is not repeated here.
  • Fig. 6 is a schematic flowchart of updating a position of a feature in an environment image according to an embodiment of the present application. As shown in FIG. 6, based on the embodiment shown in FIG. 5, the position of the update feature in the environment image based on the n + 1th frame environment image includes:
  • Step S3321 Determine the actual feature information of the features in the (n + 1) th frame environmental image by analyzing the (n + 1) th frame environmental image;
  • Step S3322 predict the prediction feature information of the features in the nth frame of the environmental image in the n + 1th frame of the environmental image according to the prediction model;
  • Step S3323 determining a first similarity between the predicted feature information and the standard feature information, and a second similarity between the actual feature information and the standard feature information;
  • Step S3324 if the first similarity is greater than or equal to the second similarity, the position of the feature in the environmental image is updated according to the predicted position of the feature in the (n + 1) th frame of the environmental image, and if the second similarity is greater than or It is equal to the first similarity, and the position of the feature in the environment image is updated according to the position of the feature in the n + 1 frame environment image.
  • the n + 1th frame of the environment image when the n + 1th frame of the environment image is collected, the n + 1th frame of the environment image may be analyzed to determine the actual feature information of the features in the n + 1th frame of the environment image.
  • the first prediction model can be obtained through machine learning training in advance.
  • the first prediction model can predict features in a certain frame of environmental image and feature information when it appears in the next frame of environmental image, thereby collecting the n + 1th
  • the first prediction model can be used to predict (for example, the feature information of the features in the first n frames of the environment image is used as an input to predict) the features in the n frame of the environment image at the n + 1 frame environment. Prediction feature information in the image.
  • the types of feature information included in the predicted feature information and the actual feature information may be the same, such as, but not limited to, shapes, colors, and relative positions with other features.
  • the environment inside the train is not static, changes may occur in some cases. For example, when collecting the nth frame of the environmental image, the features in the image are not blocked, and when collecting the n + 1th frame of the image, there are The passenger stands up and blocks the feature, which may cause the actual feature information of the feature determined through analysis of the n + 1 frame environmental image, which is not the same as the standard feature information of the feature, that is, the n frame There is a large difference in the feature information of the features in the environmental image. In this case, if the feature information in the n + 1 frame of the environment image is compared with the features in the n frame of the environment image based on the actual feature information comparison Property, you may get wrong judgment results.
  • the standard feature information is feature information about a feature stored in advance.
  • the feature is a seat in a train.
  • the shape, color, and position of the seat can be collected as a standard.
  • Feature information is stored in the robot's memory. Because the standard feature information is stored in advance, it can reflect the true feature information of the feature for later comparison with the actual feature information.
  • the predicted feature information in the n + 1th frame environmental image does not need to analyze the n + 1th frame environmental image, so as to avoid the above-mentioned errors Judge the problem.
  • the prediction result may also have errors, in order to ensure that the position of the feature in the environmental image is accurately updated, the first similarity between the predicted feature information and the standard feature information, and the actual feature information and the standard feature information may be determined. Second similarity, and compare the two.
  • the standard feature information can be measured by a feature and stored in the robot.
  • the first similarity is greater than the second similarity, it indicates that the predicted position information of the feature in the predicted n + 1 frame environmental image is more consistent with the standard position information of the feature, so that the predicted n + 1 frame The position of the feature in the environment image, and the position of the feature in the environment image is updated.
  • a second prediction model can be obtained through machine learning training in advance, and the second prediction model can predict the features in a certain frame of the environmental image and the position when it appears in the next frame of the environmental image, that is, the predicted position information.
  • the second prediction model can be used to predict (for example, the position information of the features in the first n frames of the environmental image is used as an input to predict) the features in the nth frame of the environmental image.
  • the predicted position information in the n + 1th frame of the environmental image can be obtained through machine learning training in advance, and the second prediction model can predict the features in a certain frame of the environmental image and the position when it appears in the next frame of the environmental image, that is, the predicted position information.
  • the second similarity is greater than the first similarity, it indicates that the actual feature information of the features determined by analyzing the n + 1 frame environment image is more consistent with the standard feature information of the feature, so that the n + 1 frame environment can be analyzed according to the analysis.
  • the actual position of the feature determined in the image updates the position of the feature in the environmental image.
  • the first similarity is equal to the second similarity
  • it may be classified as a case where the first similarity is greater than the second similarity, or as a case where the second similarity is greater than the first similarity.
  • Fig. 7 is a schematic flowchart illustrating whether tracking of a feature is ended according to an embodiment of the present application. As shown in FIG. 7, on the basis of the embodiment shown in FIG. 4, the determining whether the tracking of the feature has ended includes:
  • step S341 it is determined whether the feature is located in a preset area of the environment image, and if the feature is located in a preset area of the environment image, the tracking of the feature is ended.
  • the position where the robot collects environmental images will change, which will cause the feature to change in the area where the captured environmental images are located. For example, if the image acquisition device is located on the front of the robot, then When the robot moves forward, the feature moves backward relative to the robot. This movement relationship is reflected in multiple frames of the environment image. The feature generally moves from the middle of the environment image to the lower left or right of the environment image. When the robot moves When passing a certain feature, the image acquisition device cannot capture the feature, that is, the feature disappears from the environment image.
  • the feature when the feature is located in a preset area in the environment image, for example, the feature is located in the environment image In the lower left corner or the lower right corner of the center, it can be determined that the feature is about to disappear from the environment image, that is, the robot is about to pass the feature, so that it can end the tracking of the feature and determine to pass the feature.
  • the preset area can be set according to needs. For example, it can be set to the lower left corner and lower right corner of the environment image to determine whether the feature is located in the preset area of the environment image. You can select it according to your needs.
  • the position of the feature in the environmental image can be determined first, and then the distance between the position and the center of the environmental image, and the line connecting the position and the center of the environmental image with the horizontal line and vertical on the same plane as the line.
  • the angle of the line is further used to establish a coordinate system with the center of the environmental image as the origin.
  • the coordinates of the feature in the coordinate system are determined based on the distance and the included angle. Based on the coordinates, it can be determined whether the feature is located in a preset area of the environmental image.
  • the feature is a second feature, such as in a train, and the ratio of the number of the second features, such as the number of seats, to the number of the compartments is greater than two.
  • a second feature whose ratio of the number to the number of the carriages is greater than 2, and such second features may repeatedly appear in the carriage.
  • second features For example, seats, windows, etc. Taking the windows as an example, the robot will frequently pass through the windows during the movement, and there will be new windows, which may cause multiple windows to appear at the same time in the same environmental image Tracking multiple features at the same time will increase the burden on the robot to process the data. Therefore, for this type of second feature, it can be determined that when the second feature is located in a preset area in the environmental image, its end will be ended. Tracking, for example, there are 5 windows in the same environment image, then these 5 windows need to be tracked.
  • Fig. 8 is another schematic flowchart for determining whether tracking of a feature is ended according to an embodiment of the present application. As shown in FIG. 8, on the basis of the embodiment shown in FIG. 4, the determining whether the tracking of a feature is ended includes:
  • step S342 when the n + 1 frame environmental image is acquired, if the feature in the n frame environmental image does not exist in the n + 1 frame environmental image, the tracking of the feature is ended.
  • the feature is a first feature, for example, in a train, the ratio of the number of the first feature to the number of the cars is less than or equal to two.
  • first features having a ratio of the number to the number of the compartments less than or equal to 2, and such first features appear at most 2 in the compartment.
  • compartment doors, toilets, etc. Take the compartment doors as an example.
  • the robot will only pass through 2 compartment doors, which is in the same environment as the second feature such as the seat and the window.
  • two or more compartment doors do not appear in the image at the same time, so for this type of first feature, it can be determined that when it exists in the n-th frame image and does not exist in the n + 1-th frame image, When the n + 1th frame of image is collected, the tracking is ended.
  • the method is simpler, which is beneficial to reducing the load of the robot processing data.
  • this application also provides an embodiment of a positioning device in a train.
  • the embodiments of the positioning device in a train of the present application can be applied to a robot.
  • the device embodiments may be implemented by software, or by hardware or a combination of software and hardware. Taking software implementation as an example, as a logical device, it is formed by reading the corresponding computer program instructions in the non-volatile memory into the memory through the processor of the robot in which it is located.
  • FIG. 9 it is a hardware structure diagram of a robot in which a positioning device in a train is shown according to an embodiment of the present application, except for the processor, memory, network interface, and non-processor shown in FIG. 9.
  • the robot in which the device is located in the embodiment may generally include other hardware according to the actual function of the robot, and details are not described herein again.
  • Fig. 10 is a schematic block diagram of a positioning device in a train according to an embodiment of the present application.
  • the device shown in this embodiment can be applied to mobile devices, such as robots, which can move in trains, and the robots can perform distribution (e.g., equipped with a loading device on the robot), and check tickets (e.g., configured on the robot) Scanning device), cleaning (for example, installing a cleaning device on a robot), and the like.
  • the positioning device in a train may include:
  • a feature determination unit 2 is configured to determine a feature based on the environment image, wherein the feature is set in the train according to a preset rule;
  • a position determining unit 3 is configured to determine a position of the mobile device in the train based on the number of the features passed by the mobile device.
  • Fig. 11 is a schematic block diagram of a position determining unit according to an embodiment of the present application. As shown in FIG. 11, based on the embodiment shown in FIG. 10, the features include a first feature and a second feature, and the position determining unit 3 includes:
  • a carriage determining sub-unit 31, configured to determine the carriage in which the mobile device is located in the train according to the number of first features passed by the mobile device;
  • the position determining sub-unit 32 is configured to determine the position of the mobile device in the compartment where the mobile device is located according to the number of the second features that the mobile device passes in the compartment where the mobile device is located.
  • Fig. 12 is a schematic block diagram of another in-train positioning device according to an embodiment of the present application. As shown in FIG. 12, based on the embodiment shown in FIG. 11, the device further includes:
  • the relative position determining unit 4 is configured to determine a relative position between the mobile device and the second feature according to a distance between the mobile device and the second feature.
  • Fig. 13 is a schematic block diagram of another location determining unit according to an embodiment of the present application. As shown in FIG. 13, based on the embodiment shown in FIG. 11, the position determining unit 3 includes:
  • a tracking sub-unit 33 configured to track a feature according to a preset manner
  • the end determination sub-unit 34 is used to determine whether the tracking of the feature is ended, and if it is ended, it is determined to pass the feature;
  • the passing update sub-unit 35 is used to update the number of passing features.
  • Fig. 14 is a schematic block diagram of a tracking subunit according to an embodiment of the present application. As shown in FIG. 14, based on the embodiment shown in FIG. 13, the tracking sub-unit 33 includes:
  • the object determining module 331 is configured to determine whether a feature in the n-th frame of the environment image is the same as a feature in the n + 1th frame of the environment image, and n is a positive integer;
  • the position update module 332 updates the position of the feature in the environment image based on the n + 1 frame environment image if it is the same feature;
  • the tracking sub-unit 33 tracks the features in the (n + 1) th frame of the environmental image according to the preset manner.
  • Fig. 15 is a schematic block diagram of a location update module according to an embodiment of the present application. As shown in FIG. 15, based on the embodiment shown in FIG. 14, the location update module 332 includes:
  • An analysis sub-module 3321 is configured to determine actual feature information of a feature in the n + 1 frame environmental image by analyzing the n + 1 frame environmental image;
  • a prediction sub-module 3322 configured to predict features in the n-th frame environmental image according to a prediction model, and predict feature information in the n + 1-th frame environmental image;
  • a similarity submodule 3323 configured to determine a first similarity between the predicted feature information and the standard feature information, and a second similarity between the actual feature information and the standard feature information;
  • Update the submodule 3324 If the first similarity is greater than or equal to the second similarity, update the position of the feature in the environmental image according to the predicted position of the feature in the (n + 1) th frame environmental image. If the second similarity is Is greater than or equal to the first similarity, and updates the position of the feature in the environment image according to the position of the feature in the n + 1 frame environment image.
  • the ending determination subunit is configured to determine whether the second feature is located in a preset area of the environment image, and if the second feature is located in a preset area of the environment image, ending the tracking of the second feature .
  • the end determination sub-unit is configured to, when the n + 1 frame environmental image is acquired, if the feature in the n frame environmental image does not exist in the n + 1 frame environmental image, end the pair Tracking of the feature.
  • the relevant part may refer to the description of the method embodiment.
  • the device embodiments described above are only schematic, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, may be located One place, or it can be distributed across multiple network elements. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this application. Those of ordinary skill in the art can understand and implement without creative efforts.
  • An embodiment of the present application further provides an electronic device, including:
  • Memory for storing processor-executable instructions
  • the processor is configured to execute the method according to any one of the foregoing embodiments, and the electronic device may be a robot, a terminal of a controller of a driving device, or a server.
  • An embodiment of the present application further provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the steps in the method according to any one of the foregoing embodiments are performed.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Multimedia (AREA)
  • Data Mining & Analysis (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Image Analysis (AREA)
  • Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)

Abstract

一种列车内定位方法，包括：获取列车内的环境图像；基于所述环境图像确定特征物，其中，所述特征物在所述列车内按照预设规则设置；基于移动设备经过的所述特征物的数量，确定所述移动设备在所述列车中位置。

Description

列车内定位、以及室内定位 技术领域
本申请涉及定位技术领域，具体而言，涉及列车内定位以及室内定位方案。
背景技术
为了在列车内进行定位，由于车厢属于室内，应用GPS定位效果很差，因此在相关技术中，主要是通过SLAM(simultaneous localization and mapping，即时定位与地图构建)技术进行定位。
然而基于SLAM技术，一方面需要配置激光雷达扫描环境，成本较高；另一方面需要根据扫描结果构建环境地图，而对于构建的地图，需要进行滤波等图像优化过程，其中处理过程较为复杂，耗时较多，因此实时定位效果较差。
发明内容
根据本申请实施例的第一方面，提出一种列车内定位方法，包括：
获取列车内的环境图像；
基于所述环境图像确定特征物，其中，所述特征物在所述列车内按照预设规则设置；
基于移动设备经过的所述特征物的数量，确定所述移动设备在所述列车中位置。
根据本申请实施例的第二方面，提出一种列车内定位装置，包括：
图像获取单元，用于获取列车内的环境图像；
特征物确定单元，用于基于所述环境图像确定特征物，其中，所述特征物在所述列车内按照预设规则设置；
位置确定单元，用于基于移动设备经过的所述特征物的数量，确定所述移动设备在所述列车中位置。
根据本申请实施例的第三方面，提出一种移动设备，包括：
处理器；
用于存储处理器可执行指令的存储器；
其中，所述处理器被配置为执行上述的列车内定位方法。
根据本申请实施例的第四方面，提出一种计算机可读存储介质，其上存储有计算机程序，该程序被处理器执行时执行上述的列车内定位方法。
根据本申请实施例的第五方面，提出一种室内定位方法，包括：获取室内的环境图像；基于所述环境图像确定特征物，其中，所述特征物在所述室内按照预设规则设置；基于移动设备经过的所述特征物的数量，确定所述移动设备在所述室内的位置。
根据本申请实施例的第六方面，提出一种室内定位装置，包括：图像获取单元，用于获取室内的环境图像；特征物确定单元，用于基于所述环境图像确定特征物，其中，所述特征物在所述室内按照预设规则设置；位置确定单元，基于移动设备经过的所述特征物的数量，确定所述移动设备在所述室内位置。
根据本申请实施例的第七方面，提出一种移动设备，包括：处理器；用于存储处理器可执行指令的存储器；其中，所述处理器被配置为执行上述的室内定位方法。
根据本申请实施例的第八方面，提出一种计算机可读存储介质，其上存储有计算机程序，该程序被处理器执行时执行上述的室内定位方法。
根据本申请的实施例，可以根据经过的特征物的数量确定机器人在室内的位置，并且特征物是从环境图像中提取的，一方面无需通过SLAM构建地图，从而无需在机器人上设置激光雷达，因此 可以节约成本；另一方面只需确定经过的特征物的数量，相对确定机器人在地图中的坐标，计算过程简单，耗时较短，从而可以保证定位的实时性。
应当理解的是，以上的一般描述和后文的细节描述仅是示例性和解释性的，并不能限制本申请。
附图说明
此处的附图被并入说明书中并构成本说明书的一部分，示出了符合本申请的实施例，并与说明书一起用于解释本申请的原理。
图1A是根据本申请的实施例示出的一种室内定位方法的示意流程图。
图1B是根据本申请的实施例示出的一种列车内定位方法的示意流程图。
图2是根据本申请的实施例示出的一种基于移动设备经过的所述特征物的数量，确定移动设备在所述列车中位置的示意流程图。
图3是根据本申请的实施例示出的另一种列车内定位方法的示意流程图。
图4是根据本申请的实施例示出的一种确定经过的特征物的数量的示意流程图。
图5是根据本申请的实施例示出的一种根据预设方式对特征物进行跟踪的示意流程图。
图6是根据本申请的实施例示出的一种更新特征物在环境图像中的位置的示意流程图。
图7是根据本申请的实施例示出的一种确定对特征物的跟踪是否结束的示意流程图。
图8是根据本申请的实施例示出的另一种确定对特征物的跟踪是否结束的示意流程图。
图9是根据本申请的实施例示出的列车内定位装置所在机器人的一种硬件结构图。
图10是根据本申请的实施例示出的一种列车内定位装置的示意框图。
图11是根据本申请的实施例示出的一种位置确定单元的示意框图。
图12是根据本申请的实施例示出的另一种列车内定位装置的示意框图。
图13是根据本申请的实施例示出的另一种位置确定单元的示意框图。
图14是根据本申请的实施例示出的跟踪子单元的示意框图。
图15是根据本申请的实施例示出的位置更新模块的示意框图。
具体实施方式
这里将详细地对示例性实施例进行说明，其示例表示在附图中。下面的描述涉及附图时，除非另有表示，不同附图中的相同数字表示相同或相似的要素。以下示例性实施例中所描述的实施方式并不代表与本申请相一致的所有实施方式。相反，它们仅是与如所附权利要求书中所详述的、本申请的一些方面相一致的装置和方法的例子。
在本申请使用的术语是仅仅出于描述特定实施例的目的，而非旨在限制本申请。在本申请和所附权利要求书中所使用的单数形式的“一种”、“所述”和“该”也旨在包括多数形式，除非上下文清楚地表示其他含义。还应当理解，本文中使用的术语“和/或”是指并包含一个或多个相关联的列出项目的任何或所有可能组合。
应当理解，尽管在本申请可能采用术语第一、第二、第三等来描述各种信息，但这些信息不应限于这些术语。这些术语仅用来将同一类型的信息彼此区分开。例如，在不脱离本申请范围的情况下，第一信息也可以被称为第二信息，类似地，第二信息也可以被称为第一信息。取决于语境，如在此所使用的词语“如果”可以被解释成为“在……时”或“当……时”或“响应于确定”。
图1A是根据本申请的实施例示出的一种室内定位方法的示意流程图。本实施例所示的方法可以适用于移动设备，例如机器人，所述机器人可以在例如列车内的特定区域内运动。
本申请实施例的室内定位方法包括以下步骤：
步骤S1'，获取室内的环境图像；
步骤S2'，基于所述环境图像确定特征物，其中，所述特征物在所述室内按照预设规则设置；
步骤S3'，基于移动设备经过的所述特征物的数量，确定所述移动设备在所述室内的位置。
以下，以机器人在列车内移动为例对本公开的技术方案进行介绍。所述机器人可以进行配送(例如在机器人上配备载货装置)、验票(例如在机器人上配置扫描装置)、打扫(例如在机器人上配置清洁装置)等工作。
如图1B所示，所述列车内定位方法可以包括以下步骤S1至步骤S3。
步骤S1，获取列车内的环境图像。
在一个实施例中，在机器人上可以设置有图像采集设备，例如深度摄像机。在这种情况下，采集的环境图像可以是深度图像。当然，也可以根据需要选用其他图像采集设备获取环境图像。
在一个实施例中，所述列车是指具有多个车厢的车辆，例如火车、地铁、高铁、动车等。
步骤S2，基于所述环境图像确定特征物。其中，所述特征物在所述列车内按照预设规则设置。
在一个实施例中，所述特征物在所述列车内按照预设规则设置，可以包括以下一种规则或多种规则的组合：所述特征物的数量与所述列车的车厢的数量正相关，所述特征物在所述列车内按照预设顺序设置，所述特征物在所述列车内等间距设置。
在一个实施例中，对于采集到的环境图像，可以获取图像物体的bounding box，bounding box可以理解为包围图像中物体的矩形框，针对bounding box，可以根据MASK-RCNN(一种通过卷积神经网络进行物体检测算法)或者SSD(一种直接预测bounding box的坐标和类别的物体检测算法)对所获得bounding box进行处理，以确定其是否为特征物。除了上述两种方式，还可以通过机器学习训练出识别模型，并根据识别模型在环境图像中识别出特征物。具体确定特征物的过程并非本实施例的主要改进点，因此在此不再赘述。
步骤S3，基于移动设备经过的所述特征物的数量，确定所述移动设备在所述列车中位置。
在一个实施例中，特征物可以是车厢门、车窗、座椅(可以是整体的座椅，也可以是座椅的部分结构，例如座椅的把手、靠背等)、卧铺等。例如对车厢门而言，每个车厢在两端各设置有一个车厢门，按照预设规则设置可以是指，车厢门的数量与车厢的数量的比例是2，并且车厢门在列车内沿着车厢长度方向设置。又例如对于车座而言，在每个车厢中设置有K(K为正整数)排车座，按照预设规则设置可以是指，车座的数量与车厢的数量的比例是K和每排车座的数量之积，并且车座在列车内沿着车厢长度方向设置。
由于特征物在列车内按照预设规则设置，当机器人在列车内按照与预设规则相关的方式运动，那么机器人经过的特征物越多，则走过的路程越多。因此，根据走过的路程和预设规则，可以确定机器人在列车内位置。
例如按照预设规则设置是指特征物的数量与车厢的数量正相关，并且在列车内沿着车厢长度方向设置特征物。那么，在机器人沿着车厢长度方向运动时，其走过的路程就与其经过的特征物的数量正相关，因此根据机器人经过的特征物的数量，可以确定在列车中的位置。
首先可以确定机器人的起始位置和运动方向。例如，以座椅作为特征物，机器人的起始位置在车头，向车尾运动，车厢从车头向车尾为1号至8号车厢，每个车厢包含20排座椅，从车厢头到车厢尾安装了第1排至第20排座椅。若经过的特征物的数量为65，那么可以确定经过了3节车厢，当前正位于4号车厢，并且在4号车厢经过了5排座椅，当前正位于第6排座椅，从而可以确定机器人在列车中的位置是4号车厢第6排座椅。
根据本申请的实施例，可以根据经过的特征物的数量确定机器人在列车中的位置，并且特征物是从环境图像中提取的，一方面无需通过SLAM构建地图，从而无需在机器人上设置激光雷达，因此可以节约成本；另一方面只需确定经过的特征物的数量，相对于确定机器人在地图中的坐标，计算过程简单，耗时较短，从而可以保证定位的实时性。
需要说明的是，本实施例的构思也可以应用在其他空间场景，例如电影院、车间、仓库等。例如应用在电影院时，特征物可以是座椅，可以根据经过座椅的数量确定移动设备在电影院中的位置。又例如应用在车间时，可以根据经过机床的数量确定移动设备在车间中的位置。再例如应用在仓库 时，可以根据经过储藏物(例如箱子、桶)的数量确定移动设备在仓库中的位置。
图2是根据本申请的实施例示出的一种基于移动设备经过的所述特征物的数量，确定移动设备在所述列车中位置的示意流程图。如图2所示，在图1所示实施例的基础上，所述特征物包括第一特征物和第二特征物，所述基于移动设备经过的所述特征物的数量，确定移动设备在所述列车中位置，包括：
步骤S31，根据移动设备经过的第一特征物的数量，确定移动设备在所述列车中所在的车厢；
步骤S32，根据移动设备在其所在的车厢经过的第二特征物的数量，确定移动设备在其所在的车厢中的位置。
在一个实施例中，第一特征物可以是用于表征车厢的特征物。例如所述第一特征物的数量，与所述车厢的数量的比例小于或等于2，也即第一特征物在车厢中几乎不会重复出现。例如第一特征物为车厢门、卫生间等。第二特征物的数量与车厢的数量的比例大于2，也即第二特征物在车厢中会重复出现。例如第二特征物为座椅、车窗等。
虽然根据经过第二特征物的数量可以确定机器人所在的车厢，但是，一方面根据第二特征物的数据量确定机器人所在的车厢，运算量较大。例如基于图1所示的实施例，机器人需要达到的位置是4号车厢第6排座椅，为了到达该位置，所记录的经过的座椅的数量为65。
而根据本实施例，可以先根据经过的第一特征物的数量，确定机器人所在的车厢，进而根据机器人在其所在的车厢经过的第二特征物的数量，确定机器人在其所在的车厢中对应的第二特征物。例如机器人需要达到的位置是4号车厢第6排座椅，那么可以先记录经过车厢门的数量，也即只需记录经过6个车厢门(每个车厢包括2个车厢门)，就可以确定位于4号车厢了，进而再记录机器人在4号车厢所经过的座椅数量，则只需记录经过5排座椅，就可以确定机器人达到4号车厢中对应第6排座椅的位置。
相对于记录65排座椅，记录6个车厢门和5排座椅所记录的数据量更小，有利于降低机器人的运算负担。
图3是根据本申请的实施例示出的另一种列车内定位方法的示意流程图。如图3所示，在图2所示实施例的基础上，所述方法还包括：
步骤S4，根据所述移动设备与所述第二特征物的距离，确定所述移动设备与所述第二特征物的相对位置。
由于根据经过特征物的数量，只能确定机器人在列车中的位置，并且这个位置存在较大误差。例如仍以特征物为座椅为例，两排座椅的间距为1米，那么当机器人经过第n排座椅，尚未经过第n+1排座椅时，只能确定出位置是在第n排座椅和第n+1排座椅之间的区域，所以机器人确定位置的误差约为1米。
通过进一步确定机器人与第二特征物的距离，可以确定出移动设备与第二特征物的相对位置。例如获取到的环境图像为深度图像，根据座椅在环境图像中的深度，可以确定出座椅到机器人的距离，也即机器人与座椅的沿列车长度方向的相对距离。而机器人与座椅的相对距离的误差远小于两排座椅之间的距离。例如在机器人经过第n排座椅，尚未经过第n+1排座椅时，可以确定自身与第n+1排座椅的相对距离。例如相对距离为0.4米，那么可以确定机器人当前位于第n排和第n+1排座椅之间，且到第n+1排座椅0.4米的位置。这比只确定出位置是在第n排座椅和第n+1排座椅之间的区域精度要高，便于更准确地对机器人进行定位。
图4是根据本申请的实施例示出的一种确定经过的特征物的数量的示意流程图。如图4所示，在图2所示实施例的基础上，所述经过的特征物的数量通过以下方式确定包括：
步骤S33，根据预设方式对特征物进行跟踪；
步骤S34，确定对特征物的跟踪是否结束，若结束，确定经过该特征物；
步骤S35，更新经过的特征物的数量。
在一个实施例中，在环境图像中确定特征物之后，可以采用预设方式对特征物进行跟踪。其中，预设方式可以根据需要进行选择，后续实施例中主要针对两种预设方式进行示例性说明。
例如一种预设方式是在特征物在环境图像中满足特定的条件时结束对特征物的跟踪，另一种预 设方式是当特征物不存在于环境图像中(例如，从当前帧的环境图像中消失)时结束对特征物的跟踪。
每结束对一个特征物的跟踪，可以确定经过一个特征物，并更新特征物的数量。例如每经过一个特征物，在当前纪录的特征物的数量上加1。以此类推，可以确定出经过特征物的数量，根据该数量即可确定机器人的位置。
图5是根据本申请的实施例示出的一种根据预设方式对特征物进行跟踪的示意流程图。如图5所示，在图4所示实施例的基础上，所述根据预设方式对特征物进行跟踪包括：
步骤S331，确定第n帧环境图像中的特征物，与第n+1帧环境图像中的特征物是否为同一特征物，n为正整数；
步骤S332，若为同一特征物，根据第n+1帧环境图像，更新特征物在环境图像中的位置；
步骤S333，若不是同一特征物，根据所述预设方式对第n+1帧环境图像中的特征物进行跟踪。
在一个实施例中，可以连续采集多帧环境图像。在采集到第n+1帧环境图像时，可以确定第n+1帧环境图像中的特征物，并将第n+1帧环境图像中的特征物与第n帧环境图像中的特征物进行比较，例如比较特征物的bounding box。其中，可以基于NCC(Normalization cross correlation，归一化交叉相关)进行比较，并且在比较过程中还可以考虑特征物的位置(例如bounding box中心的位置)以及机器人的运动速度。例如第n帧和第n+1帧环境图像的采集时间相隔0.1秒，机器人的运动速度为0.5米/秒，而第n帧环境图像中特征物与第n+1帧环境图像中特征物的相对于机器人的位置相差1米，远大于0.05米，那么可以确定第n+1帧环境图像中特征物与第n帧环境图像中的特征物两者为不同特征物。换句话说，若第n+1帧环境图像中特征物与第n帧环境图像中特征物为同一个特征物的情况下，这两个特征物相对于机器人的位置相差大约0.05米。
除了上述方式，也可以根据需要选择其他方式进行比较，具体比较过程并非本实施例的主要改进点，因此在此不再赘述。
若比较结果确定第n+1帧环境图像中的特征物与第n帧环境图像中的特征物为同一特征物，也即该特征物在第n帧环境图像曾出现，在第n+1帧环境图像中也出现了，因此可以根据第n+1帧环境图像中的特征物的位置，更新特征物在环境图像中的位置。这样，可保证所存储的特征物的位置，与最近采集的环境图像相对应，以便后续针对每一帧环境图像准确地确定特征物在环境图像中所处的区域，进而确定出移动设备在哪一帧环境图像中经过该特征物。
若根据比较结果确定第n+1帧环境图像中的特征物与第n帧环境图像中的特征物不为同一特征物，也即该特征物在第n帧环境图像未出现，但是在第n+1帧环境图像中出现了，说明该特征物是第n+1帧环境图像中新出现的特征物，因此可以对该新出现的特征物进行跟踪，跟踪方式与对前述特征物的跟踪方式相同，在此不再赘述。
图6是根据本申请的实施例示出的一种更新特征物在环境图像中的位置的示意流程图。如图6所示，在图5所示实施例的基础上，所述基于第n+1帧环境图像，更新特征物在环境图像中的位置包括：
步骤S3321，通过分析第n+1帧环境图像，确定第n+1帧环境图像中的特征物的实际特征信息；
步骤S3322，根据预测模型，预测第n帧环境图像中的特征物在第n+1帧环境图像中的预测特征信息；
步骤S3323，确定预测特征信息与标准特征信息的第一相似度，以及实际特征信息与标准特征信息的第二相似度；
步骤S3324，若第一相似度大于或等于第二相似度，根据预测的第n+1帧环境图像中的特征物的位置，更新特征物在环境图像中的位置，若第二相似度大于或等于第一相似度，根据第n+1帧环境图像中的特征物的位置，更新特征物在环境图像中的位置。
在一个实施例中，在采集到第n+1帧环境图像时，可以对第n+1帧环境图像进行分析，从而确定第n+1帧环境图像中的特征物的实际特征信息。
另外，可以预先通过机器学习训练得到第一预测模型，第一预测模型可以预测某一帧环境图像中的特征物，在下一帧环境图像中出现时的特征信息，从而在采集到第n+1帧环境图像时，可以根 据第一预测模型，预测(例如根据前n帧环境图像中的特征物的特征信息作为输入量进行预测)第n帧环境图像中的特征物在第n+1帧环境图像中的预测特征信息。
上述预测特征信息和实际特征信息所包含的特征信息的种类可以是相同的，例如包括但不限于：形状、颜色、与其他特征物的相对位置。
由于列车内的环境并不是一成不变的，在某些情况下可能出现变化，例如在采集第n帧环境图像时，图像中的特征物未被遮挡，而在采集第n+1帧图像时，有乘客起身对特征物造成了遮挡，这就可能导致通过对第n+1帧环境图像进行分析确定的特征物的实际特征信息，与特征物的标准特征信息并不相同，也即与第n帧环境图像中特征物的特征信息存在较大差异，在这种情况下若基于实际特征信息比较确定第n+1帧环境图像中的特征物与第n帧环境图像中的特征物是否为同一特征物，可能得到错误的判断结果。
其中，标准特征信息是预先存储的有关特征物的特征信息，例如特征物为列车内的座椅，那么在机器人获取列车内的环境图像之前，可以采集座椅的形状、颜色、位置等作为标准特征信息，存储在机器人的存储器中。由于标准特征信息是预先存储的，因此可以反映特征物真实的特征信息，以供后续与实际特征信息进行对比。
而针对根据第一预测模型预测第n帧环境图像中的特征物，在第n+1帧环境图像中的预测特征信息，并无需对第n+1帧环境图像进行分析，从而可以避免上述误判问题。
进一步地，由于预测的结果也可能出现错误，为了保证准确地更新特征物在环境图像中的位置，可以确定预测特征信息与标准特征信息的第一相似度，以及实际特征信息与标准特征信息的第二相似度，并对两者进行比较。其中，标准特征信息可以通过对特征物进行测量并存储在机器人中。
若第一相似度大于第二相似度，说明预测的第n+1帧环境图像中的特征物的预测位置信息与特征物的标准位置信息更为相符，从而可以根据预测的第n+1帧环境图像中的特征物的位置，更新特征物在环境图像中的位置。
其中，可以预先通过机器学习训练得到第二预测模型，第二预测模型可以预测某一帧环境图像中的特征物，在下一帧环境图像中出现时的位置，也即预测位置信息，从而在采集到第n+1帧环境图像时，可以根据第二预测模型，来预测(例如根据前n帧环境图像中的特征物的位置信息作为输入量进行预测)第n帧环境图像中的特征物在第n+1帧环境图像中的预测位置信息。
若第二相似度大于第一相似度，说明分析第n+1帧环境图像所确定的特征物的实际特征信息与特征物的标准特征信息更为相符，从而可以根据分析第n+1帧环境图像所确定的特征物的实际位置，更新特征物在环境图像中的位置。
对于第一相似度等于第二相似度的情况，可以根据需要将其归入第一相似度大于第二相似度的情况，或者归入第二相似度大于第一相似度的情况。
图7是根据本申请的实施例示出的一种确定对特征物的跟踪是否结束的示意流程图。如图7所示，在图4所示实施例的基础上，所述确定对特征物的跟踪是否结束包括：
步骤S341，确定特征物是否位于环境图像的预设区域，其中，若特征物位于环境图像的预设区域，结束对该特征物的跟踪。
在一个实施例中，随着机器人的移动，会使得机器人采集环境图像的位置发生变化，从而导致特征物在所采集的环境图像所处的区域发生变化，例如图像采集设备位于机器人的正面，那么当机器人向前运动时，特征物相对于机器人向后运动，这种运动关系体现在多帧环境图像中，特征物一般是从环境图像的中部向环境图像中的左下或右下运动，当机器人经过某个特征物时，图像采集设备也就采集不到该特征物，也即该特征物从环境图像中消失，因此当特征物在环境图像中位于预设区域时，例如特征物位于环境图像中的左下角或右下角时，可以确定特征物即将从环境图像中消失，也即机器人即将经过该特征物，从而可以结束对该特征物的跟踪，并确定经过该特征物。
其中，预设区域可以根据需要进行设置，例如可以设置为环境图像的左下角、右下角等，确定特征物是否位于环境图像的预设区域方式，可以根据需要进行选择，以下主要介绍两种实施方式。
其一，可以先确定特征物在环境图像中的位置，然后确定该位置和环境图像中心的距离， 以及该位置和环境图像中心的连线与同该连线位于同一平面上的水平线和竖直线的夹角，进而以环境图像中心为原点建立坐标系，根据该距离和夹角确定特征物在坐标系中的坐标，进而根据该坐标可以，确定特征物是否位于环境图像的预设区域。
其二，可以通过深度学习进行确定，例如针对某一帧环境图像，在确定其中的特征物后，可以进一步确定特征物在该环境图像的特征信息，例如面积、颜色、形状等，然后基于深度学习算法对这些特征信息进行处理，从而得到特征物在环境图像中的位置，即，特征物在环境图像中的相对于环境图像中心的相对位置，进而根据该相对位置是否位于预设区域内，确定特征物是否位于环境图像的预设区域。
可选地，所述特征物为第二特征物，例如列车内，所述第二特征物的数量，例如座椅数量，与所述车厢的数量的比例大于2。
在一个实施例中，对于图5所述的实施例，可以将其应用于数量与所述车厢的数量的比例大于2的第二特征物，这类第二特征物在车厢中会重复出现，例如座椅、车窗等，以车窗为例，机器人在运动过程中会频繁地经过车窗，并且会有新的车窗，这就导致在同一张环境图像中可能同时出现多个车窗，而对多个特征物同时进行跟踪，会加重机器人处理数据的负担，因此对于这类第二特征物，可以确定当第二特征物位于环境图像中的预设区域时，就结束对其的跟踪，例如同一张环境图像中存在5个车窗，那么需要对这5个车窗进行跟踪，当其中一个车窗位于环境图像的预设区域时，虽然该车窗仍在环境图像中，但是可以结束对该车窗的跟踪，从而在接下来有新的车窗进入环境图像之前，仅需对4个车窗进行跟踪，有利于降低机器人处理数据的负担。
图8是根据本申请的实施例示出的另一种确定对特征物的跟踪是否结束的示意流程图。如图8所示，在图4所示实施例的基础上，所述确定对特征物的跟踪是否结束包括：
步骤S342，在采集到第n+1帧环境图像时，若在第n帧环境图像中的特征物，不存在于第n+1帧环境图像中，结束对该特征物的跟踪。
在一个实施例中，可以确定在第n帧环境图像中的特征物，在第n+1帧环境图像中是否存在，例如在第n+1帧环境图像中所获取到所有特征物的bounding box，与第n帧环境图像中特征物的bounding box都有较大差距，那么可以确定在第n帧环境图像中的特征物，不存在于第n+1帧环境图像中，从而结束对该特征物的跟踪。
可选地，所述特征物为第一特征物，例如在列车内，所述第一特征物的数量，与所述车厢的数量的比例小于或等于2。
在一个实施例中，对于图6所述的实施例，可以将其应用于数量与所述车厢的数量的比例小于或等于2的第一特征物，这类第一特征物在车厢至多出现2次，例如车厢门，卫生间等，以车厢门为例，机器人在运动过程中，每经过一个车厢只会经过2个车厢门，相对于座椅、车窗等第二特征物，在同一张环境图像中一般不会同时出现2个或以上车厢门，因此对于这类第一特征物，可以确定当其在第n帧图像中存在，而在第n+1帧图像中不存在时，就在采集到第n+1帧图像时结束对其的跟踪，相对于图5所示的实施例的方式较为简单，从而有利于降低机器人处理数据的负担。
与前述列车内定位方法的实施例相对应，本申请还提供了列车内定位装置的实施例。
本申请列车内定位装置的实施例可以应用在机器人上。装置实施例可以通过软件实现，也可以通过硬件或者软硬件结合的方式实现。以软件实现为例，作为一个逻辑意义上的装置，是通过其所在机器人的处理器将非易失性存储器中对应的计算机程序指令读取到内存中运行形成的。从硬件层面而言，如图9所示，为根据本申请的实施例示出的列车内定位装置所在机器人的一种硬件结构图，除了图9所示的处理器、内存、网络接口、以及非易失性存储器之外，实施例中装置所在的机器人通常根据该机器人的实际功能，还可以包括其他硬件，对此不再赘述。
图10是根据本申请的实施例示出的一种列车内定位装置的示意框图。本实施例所示的装置可以适用于移动设备，例如机器人，所述机器人可以在列车内运动，所述机器人可以进行配送(例如在机器人上配备载货装置)、验票(例如在机器人上配置扫描装置)、打扫(例如在机器人上配置清洁装置)等工作。
如图10所示，所述列车内定位装置可以包括：
图像获取单元1，用于获取列车内的环境图像；
特征物确定单元2，用于基于所述环境图像确定特征物，其中，所述特征物在所述列车内按照预设规则设置；
位置确定单元3，用于基于移动设备经过的所述特征物的数量，确定所述移动设备在所述列车中位置。
图11是根据本申请的实施例示出的一种位置确定单元的示意框图。如图11所示，在图10所示实施例的基础上，所述特征物包括第一特征物和第二特征物，所述位置确定单元3包括：
车厢确定子单元31，用于根据移动设备经过的第一特征物的数量，确定移动设备在所述列车中所在的车厢；
位置确定子单元32，用于根据移动设备在其所在的车厢经过的第二特征物的数量，确定移动设备在其所在的车厢中的位置。
图12是根据本申请的实施例示出的另一种列车内定位装置的示意框图。如图12所示，在图11所示实施例的基础上，所述装置还包括：
相对位置确定单元4，用于根据所述移动设备与所述第二特征物的距离，确定所述移动设备与所述第二特征物的相对位置。
图13是根据本申请的实施例示出的另一种位置确定单元的示意框图。如图13所示，在图11所示实施例的基础上，所述位置确定单元3包括：
跟踪子单元33，用于根据预设方式对特征物进行跟踪；
结束确定子单元34，用于确定对特征物的跟踪是否结束，若结束，确定经过该特征物；
经过更新子单元35，用于更新经过的特征物的数量。
图14是根据本申请的实施例示出的跟踪子单元的示意框图。如图14所示，在图13所示实施例的基础上，所述跟踪子单元33包括：
物体确定模块331，用于确定第n帧环境图像中的特征物，与第n+1帧环境图像中的特征物是否为同一特征物，n为正整数；
位置更新模块332，若为同一特征物，基于第n+1帧环境图像，更新特征物在环境图像中的位置；
其中，若不为同一特征物，所述跟踪子单元33根据所述预设方式对第n+1帧环境图像中的特征物进行跟踪。
图15是根据本申请的实施例示出的位置更新模块的示意框图。如图15所示，在图14所示实施例的基础上，所述位置更新模块332包括：
分析子模块3321，用于通过分析第n+1帧环境图像，确定第n+1帧环境图像中的特征物的实际特征信息；
预测子模块3322，用于根据预测模型预测第n帧环境图像中的特征物，在第n+1帧环境图像中的预测特征信息；
相似度子模块3323，用于确定预测特征信息与标准特征信息的第一相似度，以及实际特征信息与标准特征信息的第二相似度；
更新子模块3324，若第一相似度大于或等于第二相似度，根据预测的第n+1帧环境图像中的特征物的位置，更新特征物在环境图像中的位置，若第二相似度大于或等于第一相似度，根据第n+1帧环境图像中的特征物的位置，更新特征物在环境图像中的位置。
可选地，所述结束确定子单元用于确定第二特征物是否位于环境图像的预设区域，其中，若第二特征物位于环境图像的预设区域，结束对该第二特征物的跟踪。
可选地，所述结束确定子单元用于在采集到第n+1帧环境图像时，若在第n帧环境图像中的特征物，不存在于第n+1帧环境图像中，结束对该特征物的跟踪。
上述装置中各个单元、模块的功能和作用的实现过程具体详见上述方法中对应步骤的实现 过程，在此不再赘述。
对于装置实施例而言，由于其基本对应于方法实施例，所以相关之处参见方法实施例的部分说明即可。以上所描述的装置实施例仅仅是示意性的，其中所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本申请方案的目的。本领域普通技术人员在不付出创造性劳动的情况下，即可以理解并实施。
本申请的实施例还提出一种电子设备，包括：
处理器；
用于存储处理器可执行指令的存储器；
其中，所述处理器被配置为执行上述任一实施例所述的方法，所述电子设备可以是机器人，也可以是驾驶设备控制者的终端，还可以是服务器。
本申请的实施例还提出一种计算机可读存储介质，其上存储有计算机程序，该程序被处理器执行时执行上述任一实施例所述方法中的步骤。

Claims (24)

  1. 一种列车内定位方法，其特征在于，包括：
    获取列车内的环境图像；
    基于所述环境图像确定特征物，其中，所述特征物在所述列车内按照预设规则设置；
    基于移动设备经过的所述特征物的数量，确定所述移动设备在所述列车中位置。
  2. 根据权利要求1所述的方法，其特征在于，所述特征物包括第一特征物和第二特征物，所述基于所述移动设备经过的所述特征物的数量，确定所述移动设备在所述列车中位置，包括：
    根据所述移动设备经过的所述第一特征物的数量，确定所述移动设备在所述列车中所在的车厢；
    根据所述移动设备在其所在的车厢经过的所述第二特征物的数量，确定所述移动设备在其所在的车厢中的位置。
  3. 根据权利要求1所述的方法，其特征在于，所述移动设备经过的特征物的数量通过以下方式确定：
    根据预设方式对所述特征物进行跟踪；
    确定对所述特征物的跟踪是否结束，若结束，确定经过该特征物；
    更新经过的特征物的数量。
  4. 根据权利要求3所述的方法，其特征在于，所述根据预设方式对特征物进行跟踪包括：
    确定第n帧环境图像中的特征物与第n+1帧环境图像中的特征物是否为同一特征物，n为正整数；
    若为同一特征物，基于第n+1帧环境图像，更新特征物在环境图像中的位置；
    若不为同一特征物，根据所述预设方式对第n+1帧环境图像中的特征物进行跟踪。
  5. 根据权利要求4所述的方法，其特征在于，基于第n+1帧环境图像，更新特征物在环境图像中的位置包括：
    通过分析第n+1帧环境图像，确定第n+1帧环境图像中的特征物的实际特征信息；
    根据预测模型，预测第n帧环境图像中的特征物在第n+1帧环境图像中的预测特征信息；
    确定预测特征信息与标准特征信息的第一相似度，以及实际特征信息与标准特征信息的第二相似度；
    若第一相似度大于或等于第二相似度，根据预测的第n+1帧环境图像中的特征物的位置，更新特征物在环境图像中的位置，若第二相似度大于或等于第一相似度，根据第n+1帧环境图像中的特征物的位置，更新特征物在环境图像中的位置。
  6. 根据权利要求3所述的方法，其特征在于，所述确定对特征物的跟踪是否结束包括：
    确定特征物是否位于环境图像的预设区域，其中，若特征物位于环境图像的预设区域，结束对该特征物的跟踪。
  7. 根据权利要求6所述的方法，其特征在于，确定特征物是否位于环境图像的预设区域包括：
    确定所述特征物在环境图像中的位置；
    确定特征物和环境图像中心的距离，以及所述特征物在环境图像中的位置和环境图像中心的连线、与水平线的夹角，所述水平线同所述连线位于同一平面上；
    建立以环境图像中心为原点的坐标系，根据所述距离和所述夹角确定特征物在坐标系中的坐标；
    根据所述坐标，判定特征物是否位于环境图像的预设区域。
  8. 根据权利要求6所述的方法，其特征在于，确定特征物是否位于环境图像的预设区域包括：
    确定所述环境图像中的特征物；
    确定特征物在所述环境图像中的特征信息；
    根据所述特征信息，获得特征物在所述环境图像中的相对于环境图像中心的相对位置；
    根据特征物在所述环境图像中的相对位置，判定特征物是否在所述环境图像的预设区域。
  9. 根据权利要求3所述的方法，其特征在于，所述确定对特征物的跟踪是否结束包括：
    在采集到第n+1帧环境图像时，若在第n帧环境图像中的特征物，不存在于第n+1帧环境图像中，结束对该特征物的跟踪，其中n为正整数。
  10. 一种列车内定位装置，其特征在于，包括：
    图像获取单元，用于获取列车内的环境图像；
    特征物确定单元，用于基于所述环境图像确定特征物，其中，所述特征物在所述列车内按照预设规则设置；
    位置确定单元，用于基于移动设备经过的所述特征物的数量，确定所述移动设备在所述列车中位置。
  11. 一种室内定位方法，包括：
    获取室内的环境图像；
    基于所述环境图像确定特征物，其中，所述特征物在所述室内按照预设规则设置；
    基于移动设备经过的所述特征物的数量，确定所述移动设备在所述室内的位置。
  12. 根据权利要求11所述的方法，其特征在于，所述特征物包括第一特征物和第二特征物，所述基于移动设备经过的所述特征物的数量，确定移动设备在所述室内的位置，包括：
    根据移动设备经过的第一特征物的数量，确定移动设备在所述室内的与第一特征物相关的第一位置；
    根据移动设备在所述第一位置经过的第二特征物的数量，确定移动设备在室内的与第二特征物相关的第二位置。
  13. 根据权利要求11所述的方法，其特征在于，所述移动设备经过的特征物的数量通过以下方式确定：
    根据预设方式对特征物进行跟踪；
    确定对特征物的跟踪是否结束，若结束，确定经过该特征物；
    更新经过的特征物的数量。
  14. 根据权利要求13所述的方法，其特征在于，所述根据预设方式对特征物进行跟踪包括：
    确定第n帧环境图像中的特征物与第n+1帧环境图像中的特征物是否为同一特征物，n为正整数；
    若为同一特征物，基于第n+1帧环境图像，更新特征物在环境图像中的位置；
    若不为同一特征物，根据所述预设方式对第n+1帧环境图像中的特征物进行跟踪。
  15. 根据权利要求14所述的方法，其特征在于，基于第n+1帧环境图像，更新特征物在环境图像中的位置包括：
    通过分析第n+1帧环境图像，确定第n+1帧环境图像中的特征物的实际特征信息；
    根据预测模型，预测第n帧环境图像中的特征物在第n+1帧环境图像中的预测特征信息；
    确定预测特征信息与标准特征信息的第一相似度，以及实际特征信息与标准特征信息的第二相似度；
    若第一相似度大于或等于第二相似度，根据预测的第n+1帧环境图像中的特征物的位置，更新特征物在环境图像中的位置，若第二相似度大于或等于第一相似度，根据第n+1帧环境图像中的特征物的位置，更新特征物在环境图像中的位置。
  16. 根据权利要求13所述的方法，其特征在于，所述确定对特征物的跟踪是否结束包括：
    确定特征物是否位于环境图像的预设区域，其中，若特征物位于环境图像的预设区域，结束对特征物的跟踪。
  17. 根据权利要求16所述的方法，其特征在于，确定特征物是否位于环境图像的预设区域包括：
    确定所述特征物在环境图像中的位置；
    确定特征物和环境图像中心的距离，以及所述特征物在环境图像中的位置和环境图像中心的连线、与水平线的夹角，所述水平线同所述连线位于同一平面上；
    建立以环境图像中心为原点的坐标系，根据所述距离和所述夹角确定特征物在坐标系中的坐标；
    根据所述坐标，判定特征物是否位于环境图像的预设区域。
  18. 根据权利要求16所述的方法，其特征在于，确定特征物是否位于环境图像的预设区域包括：
    确定所述环境图像中的特征物；
    确定特征物在所述环境图像中的特征信息；
    根据所述特征信息，获得特征物在所述环境图像中的相对于环境图像中心的相对位置；
    根据特征物在所述环境图像中的相对位置，判定特征物是否在所述环境图像的预设区域。
  19. 根据权利要求13所述的方法，其特征在于，所述确定对特征物的跟踪是否结束包括：
    在采集到第n+1帧环境图像时，若在第n帧环境图像中的特征物，不存在于第n+1帧环境图像中，结束对该特征物的跟踪，其中n为正整数。
  20. 一种室内定位装置，其特征在于，包括：
    图像获取单元，用于获取室内的环境图像；
    特征物确定单元，用于基于所述环境图像确定特征物，其中，所述特征物在所述室内按照预设规则设置；
    位置确定单元，基于移动设备经过的所述特征物的数量，确定所述移动设备在所述室内位置。
  21. 一种移动设备，其特征在于，包括：
    处理器；
    用于存储处理器可执行指令的存储器；
    其中，所述处理器被配置为执行权利要求1至9中任一项所述的方法。
  22. 一种移动设备，其特征在于，包括：
    处理器；
    用于存储处理器可执行指令的存储器；
    其中，所述处理器被配置为执行权利要求11至19中任一项所述的方法。
  23. 一种计算机可读存储介质，其上存储有计算机程序，其特征在于，该程序被处理器执行时执行所述权利要求1至9中任一项所述方法。
  24. 一种计算机可读存储介质，其上存储有计算机程序，其特征在于，该程序被处理器执行时执行所述权利要求11至19中任一项所述方法。
PCT/CN2019/105969 2018-09-17 2019-09-16 列车内定位、以及室内定位 WO2020057462A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US17/276,823 US20210350142A1 (en) 2018-09-17 2019-09-16 In-train positioning and indoor positioning

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811082493.2A CN109195106B (zh) 2018-09-17 2018-09-17 列车内定位方法和装置
CN201811082493.2 2018-09-17

Publications (1)

Publication Number Publication Date
WO2020057462A1 true WO2020057462A1 (zh) 2020-03-26

Family

ID=64911760

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/105969 WO2020057462A1 (zh) 2018-09-17 2019-09-16 列车内定位、以及室内定位

Country Status (3)

Country Link
US (1) US20210350142A1 (zh)
CN (1) CN109195106B (zh)
WO (1) WO2020057462A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109195106B (zh) * 2018-09-17 2020-01-03 北京三快在线科技有限公司 列车内定位方法和装置
CN110068333A (zh) * 2019-04-16 2019-07-30 深兰科技（上海）有限公司 一种高铁机器人定位方法、装置及存储介质
CN110308720B (zh) * 2019-06-21 2021-02-23 北京三快在线科技有限公司 一种无人配送装置及其导航定位方法、装置

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120116676A1 (en) * 2010-11-10 2012-05-10 Gm Global Technology Operations, Inc. Method of Augmenting GPS or GPS/Sensor Vehicle Positioning Using Additional In-Vehicle Vision Sensors
US20120136505A1 (en) * 2010-11-30 2012-05-31 Aisin Aw Co., Ltd. Guiding apparatus, guiding method, and guiding program product
CN102706344A (zh) * 2012-06-08 2012-10-03 中兴通讯股份有限公司 定位处理方法及装置
CN203102008U (zh) * 2013-03-12 2013-07-31 王佳 一种餐厅服务机器人
CN105258702A (zh) * 2015-10-06 2016-01-20 深圳力子机器人有限公司 一种基于slam导航移动机器人的全局定位方法
CN105793669A (zh) * 2013-12-06 2016-07-20 日立汽车系统株式会社 车辆位置推定系统、装置、方法以及照相机装置
CN105841687A (zh) * 2015-01-14 2016-08-10 上海智乘网络科技有限公司 室内定位方法和系统
CN106289290A (zh) * 2016-07-21 2017-01-04 触景无限科技（北京）有限公司 一种路径导航系统及方法
CN109195106A (zh) * 2018-09-17 2019-01-11 北京三快在线科技有限公司 列车内定位方法和装置

Family Cites Families (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110285842A1 (en) * 2002-06-04 2011-11-24 General Electric Company Mobile device positioning system and method
DE102009015500B4 (de) * 2009-04-02 2011-01-20 Deutsches Zentrum für Luft- und Raumfahrt e.V. Verfahren und Vorrichtung zur Bestimmung einer Lageänderung eines Objekts mittels einer Stereokamera
US9749823B2 (en) * 2009-12-11 2017-08-29 Mentis Services France Providing city services using mobile devices and a sensor network
WO2011086484A1 (en) * 2010-01-12 2011-07-21 Koninklijke Philips Electronics N.V. Determination of a position characteristic for an object
KR101665386B1 (ko) * 2010-11-15 2016-10-12 한화테크윈 주식회사 로봇 위치 추정 장치 및 방법
US9020191B2 (en) * 2012-11-30 2015-04-28 Qualcomm Incorporated Image-based indoor position determination
US20150092048A1 (en) * 2013-09-27 2015-04-02 Qualcomm Incorporated Off-Target Tracking Using Feature Aiding in the Context of Inertial Navigation
CN104506857B (zh) * 2015-01-15 2016-08-17 阔地教育科技有限公司 一种摄像头位置偏离检测方法及设备
CN108241844B (zh) * 2016-12-27 2021-12-14 北京文安智能技术股份有限公司 一种公交客流统计方法、装置及电子设备
CN108181610B (zh) * 2017-12-22 2021-11-19 鲁东大学 室内机器人定位方法和系统
CN108297115B (zh) * 2018-02-02 2021-09-28 弗徕威智能机器人科技（上海）有限公司 一种机器人的自主重定位方法

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120116676A1 (en) * 2010-11-10 2012-05-10 Gm Global Technology Operations, Inc. Method of Augmenting GPS or GPS/Sensor Vehicle Positioning Using Additional In-Vehicle Vision Sensors
US20120136505A1 (en) * 2010-11-30 2012-05-31 Aisin Aw Co., Ltd. Guiding apparatus, guiding method, and guiding program product
CN102706344A (zh) * 2012-06-08 2012-10-03 中兴通讯股份有限公司 定位处理方法及装置
CN203102008U (zh) * 2013-03-12 2013-07-31 王佳 一种餐厅服务机器人
CN105793669A (zh) * 2013-12-06 2016-07-20 日立汽车系统株式会社 车辆位置推定系统、装置、方法以及照相机装置
CN105841687A (zh) * 2015-01-14 2016-08-10 上海智乘网络科技有限公司 室内定位方法和系统
CN105258702A (zh) * 2015-10-06 2016-01-20 深圳力子机器人有限公司 一种基于slam导航移动机器人的全局定位方法
CN106289290A (zh) * 2016-07-21 2017-01-04 触景无限科技（北京）有限公司 一种路径导航系统及方法
CN109195106A (zh) * 2018-09-17 2019-01-11 北京三快在线科技有限公司 列车内定位方法和装置

Also Published As

Publication number Publication date
CN109195106A (zh) 2019-01-11
CN109195106B (zh) 2020-01-03
US20210350142A1 (en) 2021-11-11

Similar Documents

Publication Publication Date Title
WO2020057462A1 (zh) 列车内定位、以及室内定位
Henein et al. Dynamic SLAM: The need for speed
US7860162B2 (en) Object tracking method and object tracking apparatus
JP7080266B2 (ja) 輸送におけるａｉベースの検査
CN104169990B (zh) 用于提供关于空闲停车位的驻车信息的方法
CN108256431B (zh) 一种手部位置标识方法及装置
US8150104B2 (en) Moving object detection apparatus and computer readable storage medium storing moving object detection program
CN110263713B (zh) 车道线检测方法、装置、电子设备及存储介质
CN110675307A (zh) 基于vslam的3d稀疏点云到2d栅格图的实现方法
JP2021165731A (ja) 測位方法、装置、コンピューティングデバイス、およびコンピュータ可読記憶媒体
CN110334569A (zh) 客流量进出识别方法、装置、设备及存储介质
CN102800102A (zh) 图像处理设备和图像处理方法
CN106295598A (zh) 一种跨摄像头目标跟踪方法及装置
CN114067295A (zh) 一种车辆装载率的确定方法、装置及车辆管理系统
CN112613424A (zh) 铁轨障碍物检测方法、装置、电子设备和存储介质
McManus et al. Distraction suppression for vision-based pose estimation at city scales
US20200034626A1 (en) Method for recognizing an object of a mobile unit
KR20190115501A (ko) 차량 인식 시스템
CN114120293A (zh) 一种地铁列车乘客检测方法及系统
US9752880B2 (en) Object linking method, object linking apparatus, and storage medium
Lumentut et al. Evaluation of recursive background subtraction algorithms for real-time passenger counting at bus rapid transit system
DE102020209054A1 (de) Vorrichtung und verfahren zur personenerkennung, -verfolgung und -identifizierung unter verwendung drahtloser signale und bilder
CN111951328A (zh) 一种物体位置检测方法、装置、设备和存储介质
CN114640807B (zh) 基于视频的对象统计方法、装置、电子设备和存储介质
Huang et al. A bus crowdedness sensing system using deep-learning based object detection

Legal Events

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

Ref document number: 19861707

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19861707

Country of ref document: EP

Kind code of ref document: A1