Labeling data obtaining method, device, equipment and storage medium

Abstract

Embodiments of the present application provide a kind of marking data obtaining method, device, equipment and storage medium.Therein, the first image material and the second image material corresponding to the same house space are obtained, the first image material includes at least one first panoramic image of house space, the second image material includes at least one second panoramic image of house space, and the first image material includes first marking data;Determine the second panoramic image matched with at least one first panoramic image respectively, obtain at least one panoramic image pair;Determine the coordinate correspondence of first target point in first panoramic image and second target point in second panoramic image in each panoramic image pair, and determine the data conversion relationship of first panoramic image and second panoramic image based on coordinate correspondence;Based on the data conversion relationship of at least one panoramic image pair, the first marking data is converted, and second marking data is obtained.The technical scheme of the embodiment of the present application improves the obtaining efficiency of marking data.

Description

Technical Field

[0001] This application relates to the field of data processing, and more particularly to a method, apparatus, device, and storage medium for obtaining labeled data. Background Technology

[0002] In the field of housing-related business, data such as the location of doors and windows, and the position of furniture can typically be labeled on panoramic images of a building to clarify its layout and configuration. Traditionally, this data is obtained by manually labeling panoramic images, which is inefficient. Summary of the Invention

[0003] This application provides a method, apparatus, device, and storage medium for obtaining labeled data, in order to improve the efficiency of obtaining labeled data.

[0004] Firstly, this application provides a method for obtaining labeled data, including:

[0005] Acquire a first image material and a second image material corresponding to the same house space. The first image material includes at least one first panoramic image of the house space, and the second image material includes at least one second panoramic image of the house space. The first image material includes first annotation data.

[0006] Each second panoramic image is determined to match the at least one first panoramic image, thereby obtaining at least one panoramic image pair;

[0007] Determine the coordinate correspondence between a first target point in the first panoramic image and a second target point in the second panoramic image in each panoramic image pair, and determine the data conversion relationship between the first panoramic image and the second panoramic image based on the coordinate correspondence.

[0008] Based on the data transformation relationship of the at least one panoramic image pair, the first annotation data is transformed to obtain the second annotation data, which is used to annotate the second image material.

[0009] Optionally, the method further includes:

[0010] Select a first target point from the first panoramic image in each panoramic image pair;

[0011] Find the second target point in the second panoramic image of each panoramic image pair that corresponds to the first target point in the first panoramic image.

[0012] Optionally, the first panoramic image includes a first ground object and a plurality of first wall objects, and the second panoramic image includes a second ground object and a plurality of second wall objects;

[0013] The step of selecting the first target point from the first panoramic image in each panoramic image pair includes:

[0014] In each panoramic image pair, the intersection point of two adjacent target first wall objects and the first ground object is selected from the first panoramic image as the first target point;

[0015] The step of finding the second target point in the second panoramic image of each panoramic image pair that corresponds to the first target point in the first panoramic image includes:

[0016] Find two target second wall objects in the second panoramic image of each panoramic image pair that correspond to the two target first wall objects in the first panoramic image;

[0017] The intersection of the two target second wall objects and the second ground object is taken as the second target point.

[0018] Optionally, the first annotation data includes the first coordinate data corresponding to each of the multiple first target points;

[0019] The step of converting the first labeled data based on the data conversion relationship of the at least one panoramic image pair to obtain the second labeled data includes:

[0020] Based on the data transformation relationship of the at least one panoramic image pair, the first coordinate data corresponding to each of the plurality of first target points are transformed to obtain the second coordinate data corresponding to each of the plurality of second target points; the second coordinate data corresponding to each of the plurality of second target points constitute the second annotation data.

[0021] Optionally, the method further includes:

[0022] The second coordinate data corresponding to each of the plurality of second target points are marked in the second image material.

[0023] Optionally, the data conversion relationship between the first panoramic image and the second panoramic image includes a conversion matrix between the first coordinate system corresponding to the first panoramic image and the second coordinate system corresponding to the second panoramic image;

[0024] The step of determining the coordinate correspondence between a first target point in the first panoramic image and a second target point in the second panoramic image in each panoramic image pair, and determining the data conversion relationship between the first panoramic image and the second panoramic image based on the coordinate correspondence, includes:

[0025] Determine the first world coordinate data of the first target point in the first panoramic image and the second world coordinate data of the second target point in the second panoramic image in each panoramic image pair;

[0026] Based on the coordinate correspondence between the first world coordinate data and the second world coordinate data in each panoramic image pair, the transformation matrix between the first coordinate system corresponding to the first panoramic image and the second coordinate system corresponding to the second panoramic image is determined.

[0027] Optionally, the first annotation data includes first coordinate data of multiple first target points in at least one first coordinate system;

[0028] The step of converting the first labeled data based on the data conversion relationship of the at least one panoramic image pair to obtain the second labeled data includes:

[0029] Based on the transformation matrix of each panoramic image pair, the first coordinate system corresponding to the first panoramic image is transformed into the second coordinate system corresponding to the second panoramic image;

[0030] The first coordinate data of multiple first target points in the first annotation data under at least one first coordinate system is converted into the second coordinate data of the corresponding second target points under the second coordinate system to obtain the second annotation data.

[0031] Optionally, the first panoramic image includes a plurality of first wall objects, and the second panoramic image includes a plurality of second wall objects;

[0032] The step of determining a second panoramic image that matches the at least one first panoramic image to obtain at least one panoramic image pair includes:

[0033] For each first panoramic image, a plurality of second wall objects with a similarity greater than a preset threshold to the plurality of first wall objects are identified, and a second panoramic image including the plurality of second wall objects is determined. The first panoramic image and the second panoramic image constitute a panoramic image pair.

[0034] Optionally, obtaining the first image material and the second image material corresponding to the same house space includes:

[0035] Acquire second image materials obtained by image acquisition of the house space;

[0036] From a plurality of pre-stored image materials, a first image material is determined that meets the similarity requirement with the second image material.

[0037] Secondly, this application provides a device for obtaining labeled data, comprising:

[0038] The acquisition module is used to acquire a first image material and a second image material corresponding to the same house space. The first image material includes at least one first panoramic image of the house space, and the second image material includes at least one second panoramic image of the house space. The first image material includes first annotation data.

[0039] The first determining module is used to determine the second panoramic image that matches the at least one first panoramic image, thereby obtaining at least one panoramic image pair.

[0040] The second determining module is used to determine the coordinate correspondence between a first target point in the first panoramic image and a second target point in the second panoramic image in each panoramic image pair, and to determine the data conversion relationship between the first panoramic image and the second panoramic image based on the coordinate correspondence.

[0041] The conversion module is used to convert the first annotation data based on the data conversion relationship of the at least one panoramic image pair to obtain second annotation data, which is used to annotate the second image material.

[0042] Optionally, the device further includes:

[0043] The selection module is used to select a first target point from the first panoramic image in each panoramic image pair; the search module is used to search for a second target point corresponding to the first target point in the first panoramic image from the second panoramic image in each panoramic image pair.

[0044] Optionally, the first panoramic image includes a first ground object and a plurality of first wall objects, and the second panoramic image includes a second ground object and a plurality of second wall objects;

[0045] The selection module is specifically used to select the intersection point of two adjacent target first wall objects and the first ground object from the first panoramic image in each panoramic image pair as the first target point;

[0046] The search module is specifically used to search for two target second wall objects in the second panoramic image of each panoramic image pair that correspond to the two target first wall objects in the first panoramic image; and to take the intersection of the two target second wall objects and the second ground object as the second target point.

[0047] Optionally, the first annotation data includes the first coordinate data corresponding to each of the multiple first target points;

[0048] The conversion module is specifically used to convert the first coordinate data corresponding to each of the plurality of first target points based on the data conversion relationship of the at least one panoramic image pair, so as to obtain the second coordinate data corresponding to each of the plurality of second target points; the second coordinate data corresponding to each of the plurality of second target points constitutes the second annotation data.

[0049] Optionally, the device further includes:

[0050] The annotation module is used to annotate the second coordinate data corresponding to each of the plurality of second target points in the second image material.

[0051] Optionally, the data conversion relationship between the first panoramic image and the second panoramic image includes a conversion matrix between the first coordinate system corresponding to the first panoramic image and the second coordinate system corresponding to the second panoramic image;

[0052] The second determining module is specifically used to determine the first coordinate data of the first target point in the first panoramic image in the world coordinate system and the second coordinate data of the second target point in the second panoramic image in the world coordinate system in each panoramic image pair; based on the coordinate correspondence between the first coordinate data and the second coordinate data in each panoramic image pair, it determines the transformation matrix between the first coordinate system corresponding to the first panoramic image and the second coordinate system corresponding to the second panoramic image.

[0053] Optionally, the first annotation data includes coordinate data of multiple first target points in at least one first coordinate system;

[0054] The conversion module is specifically used to convert the first coordinate system corresponding to the first panoramic image into the second coordinate system corresponding to the second panoramic image based on the conversion matrix of each panoramic image pair; and to convert the coordinate data of multiple first target points in the first annotation data in at least one first coordinate system into the coordinate data of the corresponding second target points in the second coordinate system to obtain the second annotation data.

[0055] Optionally, the first panoramic image includes a plurality of first wall objects, and the second panoramic image includes a plurality of second wall objects;

[0056] The first determining module is specifically used to determine, for each first panoramic image, a plurality of second wall objects whose similarity to the plurality of first wall objects is greater than a preset threshold, and a second panoramic image including the plurality of second wall objects, wherein the first panoramic image and the second panoramic image constitute a panoramic image pair.

[0057] Optionally, the acquisition module is specifically used to acquire a second image material obtained by image acquisition of the house space; and to determine a first image material from a plurality of pre-stored image materials that meets the similarity requirement with the second image material.

[0058] Thirdly, this application provides a computing device, including a storage component and a processing component; the storage component stores one or more computer program instructions, which are invoked and executed by the processing component, and the processing component executes the one or more computer program instructions to implement the annotation data acquisition method as described in the first aspect.

[0059] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a computer, implements the annotation data acquisition method as described in the first aspect.

[0060] In this embodiment, first and second image materials corresponding to the same house space can be acquired. The first image material includes at least one first panoramic image of the house space, and the second image material includes at least one second panoramic image of the house space. The first image material includes first annotation data. Second panoramic images matching at least one first panoramic image can be determined, resulting in at least one panoramic image pair. The coordinate correspondence between a first target point in the first panoramic image and a second target point in the second panoramic image in each panoramic image pair is determined. Based on this coordinate correspondence, a data conversion relationship between the first and second panoramic images is determined. Finally, based on the data conversion relationship of at least one panoramic image pair, the first annotation data is converted to obtain second annotation data, which can be used to annotate the second image material. By acquiring first and second image materials corresponding to the same house space, determining matching first and second panoramic images in the first and second image materials, at least one panoramic image pair is obtained. According to the data conversion relationship between the first and second panoramic images in each panoramic image pair, the first annotation data contained in the first panoramic image is converted into second annotation data that can be annotated in the second panoramic image, thereby automatically realizing data annotation in the second image material without manual annotation, improving data annotation efficiency.

[0061] These or other aspects of this application will become more apparent in the following description of the embodiments. Attached Figure Description

[0062] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0063] Figure 1 A flowchart of one embodiment of the annotation data acquisition method provided in this application is shown;

[0064] Figure 2 This illustration shows a structural schematic diagram of one embodiment of the annotation data acquisition device provided in this application;

[0065] Figure 3 A schematic diagram of one embodiment of a computing device provided in this application is shown. Detailed Implementation

[0066] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0067] In some of the processes described in the specification, claims, and accompanying drawings of this application, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The operation numbers, such as 101, 102, etc., are merely used to distinguish different operations and do not themselves represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a chronological order, nor do they limit "first" and "second" to different types.

[0068] The solution presented in this application is applicable to data processing scenarios, particularly in the business field related to housing, specifically in the data annotation of panoramic images corresponding to houses. By annotating data information such as the positions of doors and windows, and furniture, in the panoramic images corresponding to houses, the layout and configuration of the houses can be clearly defined. In traditional solutions, this data information is obtained by manually annotating panoramic images, which is inefficient.

[0069] To address the aforementioned technical problems, the inventors, through a series of considerations and experiments, proposed the technical solution of this application, which provides a method for obtaining labeled data. The method includes: acquiring first image material and second image material corresponding to the same house space; the first image material includes at least one first panoramic image of the house space, and the second image material includes at least one second panoramic image of the house space; the first image material includes first labeled data; determining second panoramic images that match the at least one first panoramic image to obtain at least one panoramic image pair; determining the coordinate correspondence between a first target point in the first panoramic image and a second target point in the second panoramic image in each panoramic image pair, and determining a data conversion relationship between the first panoramic image and the second panoramic image based on the coordinate correspondence; and converting the first labeled data based on the data conversion relationship of the at least one panoramic image pair to obtain second labeled data, which is used to label the second image material.

[0070] By acquiring first and second image materials corresponding to the same house space, and determining the matching first and second panoramic images in the first and second image materials, at least one panoramic image pair is obtained. Based on the data conversion relationship between the first and second panoramic images in each panoramic image pair, the first annotation data contained in the first panoramic image is converted into second annotation data that can be annotated in the second panoramic image, thereby automatically realizing the data annotation in the second image material without manual annotation, thus improving the data annotation efficiency.

[0071] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0072] like Figure 1 The diagram shown is a flowchart of one embodiment of a method for obtaining labeled data provided in this application. The method may include the following steps.

[0073] S11: Obtain the first and second image materials corresponding to the same house space.

[0074] The first image material includes at least one first panoramic image of the house space, the second image material includes at least one second panoramic image of the house space, and the first image material includes first annotation data.

[0075] The technical solution of this application is applicable to the field of data annotation in panoramic images of building spaces. A panoramic image of a building space can refer to an equidistant cylindrical projection image obtained by taking pictures of a building space at a certain position and at a certain angle using a camera.

[0076] In this embodiment, two sets of panoramic image materials corresponding to the same house space can be acquired. These two sets of panoramic image materials can refer to two sets of panoramic image materials obtained by taking pictures of the same house space from different positions and angles using a camera. Each set of panoramic image materials can include at least one panoramic image, such as a panoramic image of the living room, a panoramic image of the master bedroom, and a panoramic image of the secondary bedroom.

[0077] For ease of description, the two sets of panoramic image materials mentioned above can be referred to as the first image material and the second image material, respectively. The first image material may include first annotation data. The annotation data may include room names, door and window locations, camera shooting positions, camera shooting angles, etc., and this annotation data can be used to generate a floor plan of the house space.

[0078] Optionally, the method for obtaining the first image material and the second image material corresponding to the same house space may include:

[0079] Acquire second image materials obtained by image acquisition of the house space;

[0080] From a pool of pre-stored image materials, determine the first image material whose image similarity to the second image material meets the similarity requirement.

[0081] This involves searching a library of pre-stored image materials for a first image material that matches the second image material—that is, the first image material whose similarity meets the similarity requirements. These similarity requirements can be preset. For example, a similarity requirement could be that the panoramic images from both sets of image materials correspond to the same functional areas of the house, such as both including panoramic images of the living room, master bedroom, two secondary bedrooms, and the entryway. Another example is that a similarity requirement could be that the similarity of the furniture within the house space corresponding to the panoramic images from both sets of image materials is greater than a furniture similarity threshold. This furniture similarity threshold can be set according to the actual scenario, such as 90%, 95%, etc. Yet another example is that a similarity requirement could be that the similarity of the walls within the house space corresponding to the panoramic images from both sets of image materials is greater than a wall similarity threshold. This wall similarity threshold can also be set according to the actual scenario, and will not be elaborated further.

[0082] S12: Determine the second panoramic image that matches at least one first panoramic image to obtain at least one panoramic image pair.

[0083] Specifically, panoramic image matching can be performed based on the functional area type corresponding to the panoramic image, the furniture configured within the room, etc., to obtain at least one panoramic image pair. Of course, panoramic image matching can also be implemented in other ways, which will be described in subsequent embodiments.

[0084] S13: Determine the coordinate correspondence between the first target point in the first panoramic image and the second target point in the second panoramic image in each panoramic image pair, and determine the data conversion relationship between the first panoramic image and the second panoramic image based on the coordinate correspondence.

[0085] Optionally, before determining the coordinate correspondence, the method may further include:

[0086] Select a first target point from the first panoramic image in each panoramic image pair;

[0087] Find the second target point in the second panoramic image of each panoramic image pair that corresponds to the first target point in the first panoramic image.

[0088] Therefore, based on the coordinate correspondence between the first target point in the first panoramic image and the second target point in the second panoramic image in each panoramic image pair, the data conversion relationship between the corresponding first panoramic image and the second panoramic image can be determined.

[0089] Specifically, there are multiple ways to select the first target point and find the matching second target point, which will be described in subsequent embodiments.

[0090] S14: Based on the data transformation relationship of at least one panoramic image pair, the first annotation data is transformed to obtain the second annotation data, which can be used to annotate the second image material.

[0091] In this embodiment, first and second image materials corresponding to the same house space can be acquired. The first image material includes at least one first panoramic image of the house space, and the second image material includes at least one second panoramic image of the house space. The first image material includes first annotation data. Second panoramic images matching at least one first panoramic image can be determined, obtaining at least one panoramic image pair. The coordinate correspondence between a first target point in the first panoramic image and a second target point in the second panoramic image in each panoramic image pair is determined. Based on this coordinate correspondence, a data conversion relationship between the first and second panoramic images is determined. Finally, based on the data conversion relationship of at least one panoramic image pair, the first annotation data is converted to obtain second annotation data, which can be used to annotate the second image material. By acquiring first and second image materials corresponding to the same house space, determining matching first and second panoramic images in the first and second image materials, at least one panoramic image pair is obtained. According to the data conversion relationship between the first and second panoramic images in each panoramic image pair, the first annotation data contained in the first panoramic image is converted into second annotation data that can be annotated in the second panoramic image, thereby automatically realizing data annotation in the second image material without manual annotation, improving data annotation efficiency.

[0092] The process of matching panoramic images to obtain at least one panoramic image pair will be explained first below.

[0093] In some embodiments, the first panoramic image may include a plurality of first wall objects, and the second panoramic image may include a plurality of second wall objects.

[0094] At this point, the method for determining second panoramic images that match at least one first panoramic image to obtain at least one panoramic image pair may include:

[0095] For each first panoramic image, identify multiple second wall objects whose similarity to multiple first wall objects is greater than a preset threshold, and a second panoramic image including multiple second wall objects.

[0096] The first panoramic image and the second panoramic image can form a panoramic image pair.

[0097] Specifically, when determining the similarity of wall objects, the corner point detection model can be used first to output the coordinate data of the corresponding corner points. Then, the door and window detection model can be used to output the coordinate data of the corresponding door and window positions. Combining the coordinate data of the corner points and the coordinate data of the door and window positions, the corresponding image area can be cropped from the panoramic image as the wall object image.

[0098] Furthermore, a wall matching model can be used to obtain the distance between two wall object images in terms of features, i.e., the similarity between the two wall objects. Then, combined with a preset threshold, it can be determined whether they are the same wall object. The second panoramic image, which includes multiple second wall objects corresponding to multiple first wall objects, is combined with the first panoramic image as a panoramic image pair.

[0099] By performing wall object matching, a first panoramic image and a second panoramic image containing multiple identical wall objects are identified as panoramic image pairs. This enables panoramic image pair matching in the image material, facilitating subsequent annotation data conversion based on the panoramic image pairs.

[0100] The following describes the selection of the first target point and the process of finding the matching second target point, which can be implemented in several ways.

[0101] In some alternative implementations, the first panoramic image may include a first ground object and multiple first wall objects, and the second panoramic image may include a second ground object and multiple second wall objects.

[0102] At this point, the method for selecting the first target point from the first panoramic image in each panoramic image pair may include:

[0103] In each panoramic image pair, the intersection of two adjacent target objects, the first wall object and the first ground object, is selected as the first target point.

[0104] Furthermore, the method for finding a second target point in the second panoramic image corresponding to a first target point in the first panoramic image from each panoramic image pair may include:

[0105] Find two target second wall objects in the second panoramic image of each panoramic image pair that correspond to the two target first wall objects in the first panoramic image;

[0106] The intersection of the second wall object and the second ground object is taken as the second target point.

[0107] Among them, the method of finding two target second wall objects corresponding to two target first wall objects in the first panoramic image and matching them can be used by using the wall similarity greater than a preset threshold. This will not be elaborated further.

[0108] In some alternative implementations, the first panoramic image may include a first roof object and multiple first wall objects, and the second panoramic image may include a second roof object and multiple second wall objects.

[0109] At this point, the intersection point of two adjacent target first wall objects and a first roof object in the first panoramic image of each panoramic image pair can be selected as the first target point. Additionally, the two target second wall objects corresponding to the two target first wall objects in the matching first panoramic image in the second panoramic image of each panoramic image pair can be found, and the intersection point of the two target second wall objects and a second roof object can be selected as the second target point.

[0110] In some alternative implementations, the first panoramic image may include at least one first gate object, and the second panoramic image may include at least one second gate object. Any target vertex from any target first gate object in the first panoramic image of each panoramic image pair can be selected as a first target point. Furthermore, a target second gate object corresponding to the target first gate object in the matching first panoramic image can be found in the second panoramic image of each panoramic image pair, and the corresponding target vertex in the target second gate object can be selected as a second target point.

[0111] In some alternative implementations, the first panoramic image may include at least one first window object, and the second panoramic image may include at least one second window object. Any target vertex in any target first window object can be selected as a first target point from the first panoramic image in each panoramic image pair. Furthermore, a target second window object corresponding to the target first window object in the matching first panoramic image can be found in the second panoramic image of each panoramic image pair, and the corresponding target vertex in the target second window object can be used as a second target point.

[0112] In practical applications, there are other ways to select the target point, and this application does not impose any restrictions on this.

[0113] Once the target point is selected and determined, the coordinate correspondence between the first target point in the first panoramic image and the second target point in the second panoramic image can be determined in each panoramic image pair. Based on this coordinate correspondence, the data conversion relationship between the first panoramic image and the second panoramic image can be determined. Then, the first labeled data is converted according to the data conversion relationship.

[0114] In some embodiments, the first annotation data may include the first coordinate data corresponding to each of the multiple first target points.

[0115] At this point, the method for transforming the first annotation data to obtain the second annotation data based on the data transformation relationship of at least one panoramic image pair may include:

[0116] Based on the data transformation relationship of at least one panoramic image pair, the first coordinate data corresponding to each of the multiple first target points are transformed to obtain the second coordinate data corresponding to each of the multiple second target points.

[0117] Among them, the second coordinate data corresponding to each of the multiple second target points can constitute the second annotation data.

[0118] In some embodiments, after converting the first annotation data to obtain the second annotation data, the above method may further include:

[0119] In the second image material, label the second coordinate data corresponding to each of the multiple second target points.

[0120] In practical applications, since two sets of image materials corresponding to the same room space were obtained from different camera positions and at different angles, the coordinate systems corresponding to the panoramic images in the two sets of image materials are different. Therefore, in some embodiments, the data conversion relationship between the first panoramic image and the second panoramic image may include a conversion matrix between the first coordinate system corresponding to the first panoramic image and the second coordinate system corresponding to the second panoramic image.

[0121] At this point, the method for determining the coordinate correspondence between the first target point in the first panoramic image and the second target point in the second panoramic image in each panoramic image pair, and for determining the data conversion relationship between the first panoramic image and the second panoramic image based on the coordinate correspondence, may include:

[0122] Determine the first world coordinate data of the first target point in the first panoramic image and the second world coordinate data of the second target point in the second panoramic image in each panoramic image pair;

[0123] Based on the coordinate correspondence between the first world coordinate data and the second world coordinate data in each panoramic image pair, the transformation matrix between the first coordinate system corresponding to the first panoramic image and the second coordinate system corresponding to the second panoramic image is determined.

[0124] Furthermore, the first annotation data may include the first coordinate data of multiple first target points in at least one first coordinate system.

[0125] At this point, the method for transforming the first annotation data to obtain the second annotation data based on the data transformation relationship of at least one panoramic image pair may include:

[0126] Based on the transformation matrix of each panoramic image pair, the first coordinate system corresponding to the first panoramic image is transformed into the second coordinate system corresponding to the second panoramic image;

[0127] The first coordinate data of multiple first target points in the first annotation data under at least one first coordinate system is converted into the second coordinate data of the corresponding second target points under the second coordinate system to obtain the second annotation data.

[0128] By determining the transformation matrix for each panoramic image pair and performing coordinate system transformation according to the transformation matrix, the coordinate data of the first target point in the first coordinate system can be automatically converted into the coordinate data of the second target point in the second coordinate system, thereby obtaining the second annotation data and improving the efficiency of obtaining annotation data.

[0129] like Figure 2 The diagram shown is a structural schematic of one embodiment of a labeled data acquisition device provided in this application. The device may include the following modules.

[0130] The acquisition module 201 is used to acquire a first image material and a second image material corresponding to the same house space. The first image material includes at least one first panoramic image of the house space, and the second image material includes at least one second panoramic image of the house space. The first image material includes first annotation data.

[0131] The first determining module 202 is used to determine the second panoramic image that matches the at least one first panoramic image, thereby obtaining at least one panoramic image pair;

[0132] The second determining module 203 is used to determine the coordinate correspondence between the first target point in the first panoramic image and the second target point in the second panoramic image in each panoramic image pair, and to determine the data conversion relationship between the first panoramic image and the second panoramic image based on the coordinate correspondence.

[0133] The conversion module 204 is used to convert the first annotation data based on the data conversion relationship of the at least one panoramic image pair to obtain second annotation data, which is used to annotate the second image material.

[0134] In some embodiments, the device may further include:

[0135] A selection module is used to select a first target point from the first panoramic image in each panoramic image pair;

[0136] The search module is used to search for a second target point in the second panoramic image of each panoramic image pair that corresponds to a first target point in the first panoramic image.

[0137] In some embodiments, the first panoramic image includes a first ground object and a plurality of first wall objects, and the second panoramic image includes a second ground object and a plurality of second wall objects;

[0138] The selection module is specifically used to select the intersection point of two adjacent target first wall objects and the first ground object from the first panoramic image in each panoramic image pair as the first target point;

[0139] The search module is specifically used to search for two target second wall objects in the second panoramic image of each panoramic image pair that correspond to the two target first wall objects in the first panoramic image; and to take the intersection of the two target second wall objects and the second ground object as the second target point.

[0140] In some embodiments, the first annotation data includes first coordinate data corresponding to each of the plurality of first target points;

[0141] The conversion module is specifically used to convert the first coordinate data corresponding to each of the plurality of first target points based on the data conversion relationship of the at least one panoramic image pair, so as to obtain the second coordinate data corresponding to each of the plurality of second target points; the second coordinate data corresponding to each of the plurality of second target points constitutes the second annotation data.

[0142] In some embodiments, the apparatus further includes:

[0143] The annotation module is used to annotate the second coordinate data corresponding to each of the plurality of second target points in the second image material.

[0144] In some embodiments, the data conversion relationship between the first panoramic image and the second panoramic image includes a conversion matrix between the first coordinate system corresponding to the first panoramic image and the second coordinate system corresponding to the second panoramic image.

[0145] The second determining module is specifically used to determine the first coordinate data of the first target point in the first panoramic image in the world coordinate system and the second coordinate data of the second target point in the second panoramic image in the world coordinate system in each panoramic image pair; based on the coordinate correspondence between the first coordinate data and the second coordinate data in each panoramic image pair, it determines the transformation matrix between the first coordinate system corresponding to the first panoramic image and the second coordinate system corresponding to the second panoramic image.

[0146] In some embodiments, the first annotation data includes coordinate data of a plurality of first target points in at least one first coordinate system;

[0147] The conversion module is specifically used to convert the first coordinate system corresponding to the first panoramic image into the second coordinate system corresponding to the second panoramic image based on the conversion matrix of each panoramic image pair; and to convert the coordinate data of multiple first target points in the first annotation data in at least one first coordinate system into the coordinate data of the corresponding second target points in the second coordinate system to obtain the second annotation data.

[0148] In some embodiments, the first panoramic image includes a plurality of first wall objects, and the second panoramic image includes a plurality of second wall objects;

[0149] The first determining module is specifically used to determine, for each first panoramic image, a plurality of second wall objects whose similarity to the plurality of first wall objects is greater than a preset threshold, and a second panoramic image including the plurality of second wall objects, wherein the first panoramic image and the second panoramic image constitute a panoramic image pair.

[0150] In some embodiments, the acquisition module is specifically used to acquire a second image material obtained by image acquisition of the house space; and to determine a first image material from a plurality of pre-stored image materials that meets the similarity requirement with the second image material.

[0151] like Figure 3 The diagram shown is a structural schematic of an embodiment of a computing device provided in this application, which may include a storage component 301 and a processing component 302.

[0152] The storage component 301 can store one or more computer program instructions, which can be called and executed by the processing component 302.

[0153] Processing component 302 executes one or more of the above computer program instructions to achieve the following: Figure 1 The method for obtaining labeled data in the illustrated embodiment.

[0154] The processing component 302 may include one or more processors to execute computer instructions to complete all or part of the steps in the above-described method. Alternatively, the processing component may be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above-described method.

[0155] Storage component 301 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0156] Of course, the computing device may also include other components, such as input / output interfaces, communication components, etc.

[0157] Input / output interfaces provide interfaces between processing components and peripheral interface modules, which can be output devices, input devices, etc.

[0158] The communication components are configured to facilitate wired or wireless communication between computing devices and other devices.

[0159] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a computer, can perform the above-described functions. Figure 1 The method for obtaining labeled data in the illustrated embodiment.

[0160] This application also provides a computer program product, including a computer program that, when executed by a processor, can perform the above-described functions. Figure 1 The method for obtaining labeled data in the illustrated embodiment.

[0161] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0162] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0163] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0164] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for obtaining labeled data, characterized in that, include: Acquire a first image material and a second image material corresponding to the same house space. The first image material includes at least one first panoramic image of the house space, and the second image material includes at least one second panoramic image of the house space. The first image material includes first annotation data. The first panoramic image includes a first ground object and multiple first wall objects, and the second panoramic image includes a second ground object and multiple second wall objects. Each second panoramic image is determined to match the at least one first panoramic image, thereby obtaining at least one panoramic image pair; In each panoramic image pair, the intersection point of two adjacent target first wall objects and the first ground object is selected from the first panoramic image as the first target point; In each panoramic image pair, find two target second wall objects in the second panoramic image that correspond to the two target first wall objects in the first panoramic image, and take the intersection of the two target second wall objects and the second ground object as the second target point; Determine the coordinate correspondence between a first target point in the first panoramic image and a second target point in the second panoramic image in each panoramic image pair, and determine the data conversion relationship between the first panoramic image and the second panoramic image based on the coordinate correspondence. Based on the data transformation relationship of the at least one panoramic image pair, the first annotation data is transformed to obtain the second annotation data, which is used to annotate the second image material.

2. The method according to claim 1, characterized in that, The method further includes: Select the first target point from the first panoramic image in each panoramic image pair; Find the second target point in the second panoramic image of each panoramic image pair that corresponds to the first target point in the first panoramic image.

3. The method according to claim 1, characterized in that, The first annotation data includes the first coordinate data corresponding to each of the multiple first target points; The step of converting the first labeled data based on the data conversion relationship of the at least one panoramic image pair to obtain the second labeled data includes: Based on the data transformation relationship of the at least one panoramic image pair, the first coordinate data corresponding to each of the plurality of first target points are transformed to obtain the second coordinate data corresponding to each of the plurality of second target points; the second coordinate data corresponding to each of the plurality of second target points constitute the second annotation data.

4. The method according to claim 3, characterized in that, The method further includes: The second coordinate data corresponding to each of the plurality of second target points are marked in the second image material.

5. The method according to claim 3, characterized in that, The data conversion relationship between the first panoramic image and the second panoramic image includes the conversion matrix between the first coordinate system corresponding to the first panoramic image and the second coordinate system corresponding to the second panoramic image; The step of determining the coordinate correspondence between a first target point in the first panoramic image and a second target point in the second panoramic image in each panoramic image pair, and determining the data conversion relationship between the first panoramic image and the second panoramic image based on the coordinate correspondence, includes: Determine the first world coordinate data of the first target point in the first panoramic image and the second world coordinate data of the second target point in the second panoramic image in each panoramic image pair; Based on the coordinate correspondence between the first world coordinate data and the second world coordinate data in each panoramic image pair, the transformation matrix between the first coordinate system corresponding to the first panoramic image and the second coordinate system corresponding to the second panoramic image is determined.

6. The method according to claim 5, characterized in that, The first annotation data includes the first coordinate data of multiple first target points in at least one first coordinate system; The step of converting the first labeled data based on the data conversion relationship of the at least one panoramic image pair to obtain the second labeled data includes: Based on the transformation matrix of each panoramic image pair, the first coordinate system corresponding to the first panoramic image is transformed into the second coordinate system corresponding to the second panoramic image; The first coordinate data of multiple first target points in the first annotation data under at least one first coordinate system is converted into the second coordinate data of the corresponding second target points under the second coordinate system to obtain the second annotation data.

7. The method according to claim 1, characterized in that, The first panoramic image includes multiple first wall objects, and the second panoramic image includes multiple second wall objects; The step of determining a second panoramic image that matches the at least one first panoramic image to obtain at least one panoramic image pair includes: For each first panoramic image, a plurality of second wall objects with a similarity greater than a preset threshold to the plurality of first wall objects are identified, and a second panoramic image including the plurality of second wall objects is determined. The first panoramic image and the second panoramic image constitute a panoramic image pair.

8. The method according to claim 1, characterized in that, The acquisition of the first and second image materials corresponding to the same room space includes: Acquire second image materials obtained by image acquisition of the house space; From a plurality of pre-stored image materials, a first image material is determined that meets the similarity requirement with the second image material.

9. A device for obtaining labeled data, characterized in that, include: The acquisition module is used to acquire a first image material and a second image material corresponding to the same house space. The first image material includes at least one first panoramic image of the house space, and the second image material includes at least one second panoramic image of the house space. The first image material includes first annotation data. The first panoramic image includes a first ground object and a plurality of first wall objects, and the second panoramic image includes a second ground object and a plurality of second wall objects. The first determining module is used to determine the second panoramic image that matches the at least one first panoramic image, thereby obtaining at least one panoramic image pair. The selection module is used to select the intersection point of two adjacent target first wall objects and the first ground object from the first panoramic image in each panoramic image pair as the first target point; The search module is used to search for two target second wall objects in the second panoramic image of each panoramic image pair that correspond to two target first wall objects in the first panoramic image, and to take the intersection of the two target second wall objects and the second ground object as the second target point; The second determining module is used to determine the coordinate correspondence between a first target point in the first panoramic image and a second target point in the second panoramic image in each panoramic image pair, and to determine the data conversion relationship between the first panoramic image and the second panoramic image based on the coordinate correspondence. The conversion module is used to convert the first annotation data based on the data conversion relationship of the at least one panoramic image pair to obtain second annotation data, which is used to annotate the second image material.

10. A computing device, characterized in that, It includes a storage component and a processing component; the storage component stores one or more computer program instructions, which are invoked and executed by the processing component, and the processing component executes the one or more computer program instructions to implement the annotation data acquisition method as described in any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, The device contains a computer program that, when executed by a computer, implements the annotation data acquisition method as described in any one of claims 1 to 8.

Images