A 360 panorama image-oriented labeling method and system

Abstract

The application provides a kind of marking method and system for 360 panoramic image, the method comprises: obtaining multiple single views covering full circumferential direction, splicing multiple single views to obtain panoramic image, calculating the coordinate conversion parameters between panoramic image and each single view, and writing corresponding coordinate conversion formula;Obtain 3D click coordinates when marking panoramic image, convert 3D click coordinates into map coordinates in panoramic image, map coordinates to each single view according to the coordinate conversion parameters;In the panoramic image, the current marking position and the next marking position are connected into a line in turn, until all the mapping marking lines in each single view are obtained.The application collects multiple single views in different directions and splices into panoramic image, only needs to mark once in panoramic image, can obtain the marking result of each single view according to the mapping relationship between panoramic image and each single view, greatly improves the marking efficiency and saves the computing power.

Description

Technical Field

[0001] This invention relates to the field of road labeling technology, and more specifically, to a labeling method and system for 360-degree panoramic images. Background Technology

[0002] With the rapid development of autonomous driving, the demand for road sign classification is increasing, leading to a greater need for computing power to meet these requirements. At the same time, as the requirements for labeling efficiency become increasingly stringent, how to design a system that balances the growing computing power consumption with labeling efficiency to adapt to the development of autonomous driving is a topic worthy of research. Summary of the Invention

[0003] This invention addresses the technical problems existing in the prior art by providing a method and system for labeling 360-degree panoramic images. It collects multiple single-path views from different directions and stitches them into a panoramic image. Only one labeling is required in the panoramic image, and the labeling results of each single-path view can be obtained based on the mapping relationship between the panoramic image and each single-path view, which greatly improves the labeling efficiency.

[0004] According to a first aspect of the present invention, a method for annotation of 360-degree panoramic images is provided, comprising:

[0005] Obtain multiple single-path views covering the entire circumference, stitch the multiple single-path views to obtain a panoramic view, calculate the coordinate transformation parameters between the panoramic view and each single-path view, and write the corresponding coordinate transformation formulas;

[0006] Obtain the 3D click coordinates when annotating the panoramic icon, convert the 3D click coordinates into texture coordinates in the panoramic image, and map the texture coordinates to each single-path view according to the coordinate conversion parameters; in the panoramic image, connect the current annotation position with the next annotation position to form a line until all the mapped annotation lines in each single-path view are obtained.

[0007] Based on the above technical solution, the present invention can also be improved as follows.

[0008] Optionally, obtaining multiple single-path views covering the entire circumference and stitching the multiple single-path views to obtain a panoramic image includes:

[0009] With the shooting point as the origin, multiple single-path views are sequentially acquired along the circumference of the shooting point, so that the field of view of all single-path views is superimposed to cover the entire circumference of the shooting point;

[0010] By stitching together all the single-view images in the order they were acquired, a panoramic view of the entire shooting point can be obtained.

[0011] Optionally, the plurality of single-path views include at least the front view, left front view, right front view, left side view, right side view, left rear view, and right rear view with the shooting point as the origin.

[0012] Optionally, the calculation of coordinate transformation parameters between the panoramic image and each individual view, and the writing of corresponding coordinate transformation formulas, includes:

[0013] Set the panoramic coordinates x i Convert to single-path view coordinates x r The formula is x r =(PM) -1 x i ,

[0014] Where P represents the intrinsic and extrinsic parameters of the projection matrix camera, and M represents the 3*3 coordinate transformation matrix for mapping the panoramic image to a single-path view;

[0015] Assume (x) o y o Let (x, y) be the coordinates in the panoramic view, and (x, y) be the coordinates in the single-view view. Then, the panoramic coordinates (x, y) are... o y o The conversion formula from (x, y) coordinates in a single-path view to (x, y) coordinates is:

[0016] x=((M 22 -M 32 y o (M) 33 x o -M 13 )-(M 12 -M 32 x o (M) 33 y o -M 23 )) / ((M 22 -M 32 y o

[0017] (M) 11 -M 31 x o )-(M 12 -M 32 x o (M) 21 -M 31 y o )),

[0018] y=((M 21 -M 31 y o (M) 33 x o -M 13)-(M 11 -M 31 x o (M) 33 y o -M 23 )) / ((M 21 -M 31 y o

[0019] (M) 12 -M 32 x o )-(M 11 -M 31 x o (M) 22 -M 32 y o ));

[0020] Among them, M 11 M represents the values ​​in the first row and first column of a 3x3 coordinate transformation matrix M. 12 The values ​​in the first row and second column of the 3x3 coordinate transformation matrix M are the values ​​of M. 13 M 21 M 22 M 23 M 31 M 32 M 33 And so on.

[0021] Optionally, obtaining the 3D click coordinates when annotating the panoramic image and converting the 3D click coordinates into texture coordinates in the panoramic image includes:

[0022] Load and display the 3D panoramic image. Click on the 3D panoramic image at the current annotation location to add an annotation.

[0023] Obtain the 3D click coordinates (x', y', z') of the marked point, and calculate the texture coordinates (u, v) on the panoramic image corresponding to the 3D click coordinates (x', y', z') according to the formula for converting sphere coordinates to texture coordinates;

[0024] Based on the directional differences between each single-view, the coordinate transformation parameters between the panoramic image and each single-view are set respectively. The texture coordinates (u, v) are used as the panoramic image coordinates (xo, yo) and substituted into the coordinate transformation formula between the panoramic image and each single-view to obtain the mapping coordinates (x1, y1) from the texture coordinates (u, v) to the single-view.

[0025] Continue labeling to obtain the mapped coordinates (x2, y2) of the next label position in the single-path view.

[0026] Optionally, the step of calculating the texture coordinates (u, v) on the panoramic image corresponding to the 3D click coordinates (x', y', z') based on the spherical coordinate to texture coordinate formula includes:

[0027] Substitute the values ​​of the 3D click coordinates (x', y', z') into the formula for converting sphere coordinates to texture coordinates to calculate the texture coordinates (u, v) on the panoramic image;

[0028] The formula for the sphere coordinate transfer map coordinates is:

[0029] u = arctan(z / x) / π / 2*a

[0030] v=b-(arcsin(y / radius) / π+0.5)*b

[0031] Where radius is the radius of the 3D sphere, which can be set to an empirical value, and a*b is the resolution of the panoramic image.

[0032] Optionally, the step of sequentially connecting the current annotation position with the next annotation position in the panoramic image until all the mapped annotation lines in each single-path view are obtained includes:

[0033] In the panoramic image, connect the current annotation position with the next annotation position to form a line, and obtain the annotated line segment in the panoramic image;

[0034] Traverse all the marked locations in the panoramic image until all the marked lines in the panoramic image are drawn;

[0035] Based on the mapping relationship between the panoramic view and each individual view, all mapping annotation lines in each individual view are obtained.

[0036] According to a second aspect of the present invention, a labeling system for 360° panoramic images is provided, comprising:

[0037] The conversion module is used to acquire multiple single-path views covering the entire circumference, stitch the multiple single-path views to obtain a panoramic image, and also to calculate the coordinate transformation parameters between the panoramic image and each single-path view and write the corresponding coordinate transformation formulas.

[0038] The annotation module is used to obtain the 3D click coordinates when annotating the panoramic image, convert the 3D click coordinates into texture coordinates in the panoramic image, and also to map the texture coordinates to each single-path view according to the coordinate conversion parameters; in the panoramic image, the current annotation position and the next annotation position are connected to form a line until all the mapped annotation lines in each single-path view are obtained.

[0039] According to a third aspect of the present invention, an electronic device is provided, including a memory and a processor, wherein the processor is configured to implement the steps of the above-described annotation method for 360-degree panoramic images when executing a computer management program stored in the memory.

[0040] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, on which a computer management class program is stored, which, when executed by a processor, implements the steps of the above-described annotation method for 360-degree panoramic images.

[0041] This invention provides a method, system, electronic device, and storage medium for annotating 360° panoramic images. It utilizes a 360° stereoscopic scene, acquiring multiple single-path views from different directions and stitching them together to form a panoramic image covering the entire stereoscopic scene. Only one annotation is required within the stereoscopic scene; the annotation results for each single-path view are obtained in real time based on the mapping relationship between the panoramic image and each single-path view. This significantly improves the efficiency of sequential annotation of single-path views and saves computational resources. This invention integrates repetitive annotation work on the same scene, consolidating fragmented tasks and saving computational resources. It enables the annotation of a 3D stereoscopic scene in a single step, completing the annotation of multiple single-path views, thus increasing annotation efficiency many times over compared to traditional single-path planar annotation, greatly enhancing annotation efficiency. Attached Figure Description

[0042] Figure 1 A flowchart of a labeling method for 360° panoramic images provided by the present invention;

[0043] Figure 2 This is a schematic diagram illustrating the mapping relationship between the panoramic view and each individual view in a specific embodiment of the present invention;

[0044] Figure 3 A block diagram of a labeling system for 360-degree panoramic images provided by the present invention;

[0045] Figure 4 A schematic diagram of the hardware structure of a possible electronic device provided by the present invention;

[0046] Figure 5 This is a schematic diagram of the hardware structure of a possible computer-readable storage medium provided by the present invention. Detailed Implementation

[0047] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0048] Figure 1 A flowchart of a labeling method for 360° panoramic images provided by this invention is shown below. Figure 1 As shown, the method includes:

[0049] S1. Obtain multiple single-path views covering the entire circumference, stitch the multiple single-path views to obtain a panoramic view, calculate the coordinate transformation parameters between the panoramic view and each single-path view, and write the corresponding coordinate transformation formula.

[0050] S2, obtain the 3D click coordinates when annotating the panoramic icon, convert the 3D click coordinates into texture coordinates in the panoramic image, and map the texture coordinates to each single-path view according to the coordinate conversion parameters; in the panoramic image, connect the current annotation position with the next annotation position to form a line until all the mapped annotation lines in each single-path view are obtained.

[0051] Understandably, given the shortcomings in the background technology, this invention proposes a labeling method for 360° panoramic images. It employs a 360° stereoscopic scene, acquires multiple single-path views from different directions, and stitches them together to form a panoramic image covering the entire stereoscopic scene. Only one labeling operation is required in the stereoscopic scene, and the labeling results for each single-path view can be obtained in real time based on the mapping relationship between the panoramic image and each single-path view. This significantly improves the labeling efficiency of sequentially labeling single-path views and saves computational power. This invention integrates repetitive labeling work for the same scene, consolidating fragmented tasks and saving computational power; it achieves one-time labeling of a 3D stereoscopic scene, completing all single-path labeling operations, thus increasing labeling efficiency many times over compared to the original single-path planar labeling, greatly improving labeling efficiency.

[0052] In one possible embodiment, step S1, which involves acquiring multiple single-path views covering the entire circumference and stitching the multiple single-path views together to obtain a panoramic view, includes:

[0053] With the shooting point as the origin, multiple single-path views are sequentially acquired along the circumference of the shooting point, so that the field of view of all single-path views is superimposed to cover the entire circumference of the shooting point;

[0054] By stitching together all the single-view images in the order they were acquired, a panoramic view of the entire shooting point can be obtained.

[0055] Understandably, in one embodiment, the vehicle-mounted camera captures individual road views. Specifically, taking the vehicle as the origin, it rotates clockwise in seven directions to capture road view photos. These seven road view photos are then stitched together from a clockwise perspective to create a panoramic image with a resolution of 10430*5215. The resulting panoramic image encompasses the vehicle's 360-degree circumferential view.

[0056] In one possible embodiment, the plurality of single-path views include at least a front view, a left front view, a right front view, a left side view, a right side view, a left rear view, and a right rear view, with the shooting point as the origin.

[0057] It is understandable that during vehicle driving, the markers behind the vehicle (such as road signs) are relatively weakened, so the markers behind the vehicle can be appropriately weakened in the panoramic view.

[0058] In one possible embodiment, step S1, which involves calculating the coordinate transformation parameters between the panoramic view and each individual view, and writing the corresponding coordinate transformation formula, includes:

[0059] Set the panoramic coordinates x i Convert to single-path view coordinates x r The formula is x r =(PM) -1 x i ,

[0060] Where P represents the intrinsic and extrinsic parameters of the projection matrix camera, and M represents the 3*3 coordinate transformation matrix for mapping the panoramic image to a single-path view;

[0061] Assume (x) o y o Let (x, y) be the coordinates in the panoramic view, and (x, y) be the coordinates in the single-view view. Then, the panoramic coordinates (x, y) are... o y o The conversion formula from (x, y) coordinates in a single-path view to (x, y) coordinates is:

[0062] x=((M 22 -M 32 y o (M) 33 x o -M 13 )-(M 12 -M 32 x o (M) 33 y o -M 23 )) / ((M 22 -M 32 y o

[0063] (M) 11 -M 31 x o )-(M 12 -M 32 x o (M) 21 -M 31 y o )),

[0064] y=((M 21 -M 31 y o (M) 33 x o -M13 )-(M 11 -M 31 x o (M) 33 y o -M 23 )) / ((M 21 -M 31 y o

[0065] (M) 12 -M 32 x o )-(M 11 -M 31 x o (M) 22 -M 32 y o ));

[0066] Among them, M 11 M represents the values ​​in the first row and first column of a 3x3 coordinate transformation matrix M. 12 The values ​​in the first row and second column of the 3x3 coordinate transformation matrix M are the values ​​of M. 13 M 21 M 22 M 23 M 31 M 32 M 33 And so on.

[0067] It is understandable that P represents the intrinsic and extrinsic parameters of the projection matrix camera, which includes the angles captured by the camera in each individual view. Therefore, the parameter P corresponding to each individual view is different. Thus, when converting the panoramic image to each individual view, although the same coordinate transformation formula is used, the difference in the coordinate transformation parameters in the formula allows for a one-to-one mapping of the coordinates in the panoramic image to each individual view with high accuracy.

[0068] In one possible embodiment, step S2, which involves obtaining the 3D click coordinates when annotating the panoramic image and converting the 3D click coordinates into texture coordinates in the panoramic image, includes:

[0069] The web front-end uses the 3D framework PANOLENS to load and display 3D panoramic images. The operator can click on the 3D panoramic image at the current annotation location to make annotations.

[0070] Obtain the 3D click coordinates (x', y', z') of the marked point, and calculate the texture coordinates (u, v) on the panoramic image corresponding to the 3D click coordinates (x', y', z') according to the formula for converting sphere coordinates to texture coordinates;

[0071] Based on the directional differences between each single-view, the coordinate transformation parameters between the panoramic image and each single-view are set respectively. The texture coordinates (u, v) are used as the panoramic image coordinates (xo, yo) and substituted into the coordinate transformation formula between the panoramic image and each single-view to obtain the mapping coordinates (x1, y1) from the texture coordinates (u, v) to the single-view.

[0072] Continue labeling to obtain the mapped coordinates (x2, y2) of the next label position in the single-path view.

[0073] It is understandable that this embodiment labels adjacent positions on a point-by-point basis, thus obtaining the endpoints of each line segment to be labeled. After labeling the current position, the next position can be labeled according to the labeling plan, and the previously labeled next position is used as the new current position, thereby enabling the labeling of all positions in the panoramic image.

[0074] In one possible embodiment, step S2, which involves calculating the texture coordinates (u, v) on the panoramic image corresponding to the 3D click coordinates (x', y', z') based on the spherical coordinate to texture coordinate formula, includes:

[0075] Substitute the values ​​of the 3D click coordinates (x', y', z') into the formula for converting sphere coordinates to texture coordinates to calculate the texture coordinates (u, v) on the panoramic image;

[0076] The formula for the sphere coordinate transfer map coordinates is:

[0077] u = arctan(z / x) / π / 2*a

[0078] v=b-(arcsin(y / radius) / π+0.5)*b

[0079] Wherein, radius is the radius of the 3D sphere, which can be set to an empirical value; in this embodiment, the default radius is 5000; a*b is the resolution of the panoramic image, for example, a resolution of 10432*5215.

[0080] It is understandable that this embodiment uses the formula of converting spherical coordinates to texture coordinates to convert the coordinates of the annotation points in the three-dimensional scene into two-dimensional texture coordinates, which makes it easier for subsequent steps to map the texture coordinates to the single-path view and obtain the two-dimensional annotation results in the single-path view.

[0081] In one possible embodiment, step S2, which involves sequentially connecting the current annotation position with the next annotation position in the panoramic view until all the mapped annotation lines in each single-path view are obtained, includes:

[0082] In the panoramic image, connect the current annotation position with the next annotation position to form a line, and obtain the annotated line segment in the panoramic image;

[0083] Traverse all the marked locations in the panoramic image until all the marked lines in the panoramic image are drawn;

[0084] Based on the mapping relationship between the panoramic view and each individual view, all mapping annotation lines in each individual view are obtained.

[0085] Understandably, after obtaining adjacent labeled locations in the panoramic image, spatial connections can be made between these adjacent locations to draw the labeled line segments between them. This embodiment first labels adjacent locations as points, and then draws several labeled line segments by connecting these points to form multiple polylines, thus labeling road markings such as lane lines, pedestrians, vehicles, and traffic signs. Based on the mapping relationship between the panoramic image and each individual road view mentioned in the aforementioned embodiment, the mapped labeled lines in each individual road view can be obtained in real time, achieving the goal of obtaining multiple labeled results with a single labeling, greatly saving computing power and improving labeling efficiency.

[0086] Figure 3 This invention provides a structural diagram of a labeling system for 360° panoramic images, as shown in the embodiment of the invention. Figure 3 As shown, a labeling system for 360° panoramic images includes a conversion module and a labeling module, wherein:

[0087] The conversion module is used to acquire multiple single-path views covering the entire circumference, stitch the multiple single-path views to obtain a panoramic image, and also to calculate the coordinate transformation parameters between the panoramic image and each single-path view and write the corresponding coordinate transformation formulas.

[0088] The annotation module is used to obtain the 3D click coordinates when annotating the panoramic image, convert the 3D click coordinates into texture coordinates in the panoramic image, and map the texture coordinates to each single-path view according to the coordinate conversion parameters; it is also used to connect the current annotation position with the next annotation position in the panoramic image sequentially until all the mapped annotation lines in each single-path view are obtained.

[0089] It is understood that the annotation system for 360-degree panoramic images provided by this invention corresponds to the annotation methods for 360-degree panoramic images provided in the foregoing embodiments. The relevant technical features of the annotation system for 360-degree panoramic images can be referred to the relevant technical features of the annotation methods for 360-degree panoramic images, and will not be repeated here.

[0090] Please see Figure 4 , Figure 4 This is a schematic diagram illustrating an embodiment of the electronic device provided in this invention. For example... Figure 4As shown, an embodiment of the present invention provides an electronic device 400, including a memory 410, a processor 420, and a computer program 411 stored in the memory 410 and executable on the processor 420. When the processor 420 executes the computer program 411, it performs the following steps:

[0091] Obtain multiple single-path views covering the entire circumference, stitch the multiple single-path views to obtain a panoramic view, calculate the coordinate transformation parameters between the panoramic view and each single-path view, and write the corresponding coordinate transformation formulas;

[0092] Obtain the 3D click coordinates when annotating the panoramic icon, convert the 3D click coordinates into texture coordinates in the panoramic image, and map the texture coordinates to each single-path view according to the coordinate conversion parameters; in the panoramic image, connect the current annotation position with the next annotation position to form a line until all the mapped annotation lines in each single-path view are obtained.

[0093] Please see Figure 5 , Figure 5 This is a schematic diagram illustrating an embodiment of a computer-readable storage medium provided by the present invention. (See diagram below.) Figure 5 As shown, this embodiment provides a computer-readable storage medium 500 on which a computer program 511 is stored. When the computer program 511 is executed by a processor, it performs the following steps:

[0094] Obtain multiple single-path views covering the entire circumference, stitch the multiple single-path views to obtain a panoramic view, calculate the coordinate transformation parameters between the panoramic view and each single-path view, and write the corresponding coordinate transformation formulas;

[0095] Obtain the 3D click coordinates when annotating the panoramic icon, convert the 3D click coordinates into texture coordinates in the panoramic image, and map the texture coordinates to each single-path view according to the coordinate conversion parameters; in the panoramic image, connect the current annotation position with the next annotation position to form a line until all the mapped annotation lines in each single-path view are obtained.

[0096] This invention provides a method, system, and storage medium for annotating 360° panoramic images. It utilizes a 360° stereoscopic scene, acquiring multiple single-path views from different directions and stitching them together to form a panoramic image covering the entire stereoscopic scene. Only one annotation is needed within the stereoscopic scene; the annotation results for each single-path view can be obtained in real-time based on the mapping relationship between the panoramic image and each single-path view. This significantly improves the efficiency of sequential annotation of single-path views and saves computational power. This invention integrates repetitive annotation work for the same scene, consolidating fragmented tasks and saving computational power. It achieves one-time annotation of a 3D stereoscopic scene, completing annotation of multiple single-path views, thus increasing annotation efficiency many times over compared to traditional single-path planar annotation, greatly enhancing annotation efficiency.

[0097] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0098] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0099] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0100] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0101] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0102] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0103] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A labeling method for a 360 panorama image, characterized by, The method comprises the following steps: acquiring a plurality of single-path views covering full circumferential direction, splicing the plurality of single-path views to obtain a panoramic view, calculating coordinate conversion parameters between the panoramic view and each single-path view, and compiling corresponding coordinate conversion formulas; the method comprises the following steps: Set panorama map coordinate x i Turn single road view coordinate x r The formula is , wherein P is a projection matrix of camera intrinsic and extrinsic parameters, M is a 3 3 coordinate conversion matrix Assume (x o , y o ) is the coordinate in the panorama, (x, y) is the coordinate in the single view, then the conversion formula from panorama coordinate (x o , y o ) to single view coordinate (x, y) is: x = ((M 22 M 32 y o )(M 33 x o M 13 ) (M 12 M 32 x o )(M 33 y o M 23 )) / ((M 22 M 32 y o )(M 11 M 31 x o ) (M 12 M 32 x o )(M 21 M 31 y o )) y = ((M 21 M 31 y o )(M 33 x o M 13 ) (M 11 M 31 x o )(M 33 y o M 23 )) / ((M 21 M 31 y o )(M 12 M 32 x o ) (M 11 M 31 x o )(M 22 M 32 y o )) wherein M 11 is 3 3 the value of the coordinate conversion matrix M in the first row and the first column, M 12 is 3 3 the value of the coordinate conversion matrix M in the first row and the second column, M 13 , M 21 , M 22 , M 23 , M 31 , M 32 , M 33 and so on; acquiring 3D click coordinates when labeling the panoramic view, converting the 3D click coordinates into texture coordinates in the panoramic view, and mapping the texture coordinates to each single-path view according to the coordinate conversion parameters; the method comprises the following steps: loading and displaying the 3D panoramic view, and labeling the 3D panoramic view by clicking at a current labeling position; acquiring 3D click coordinates (x', y', z') of a labeled point, and calculating texture coordinates (u, v) on the panoramic view corresponding to the 3D click coordinates (x', y', z') according to a spherical coordinate to texture coordinate conversion formula; Based on the directional differences between each individual view, coordinate transformation parameters are set between the panoramic image and each individual view, and the texture coordinates (u, v) are used as the panoramic image coordinates (x, v). o y o Substituting the coordinate transformation formula between the panoramic image and each single-path view, we obtain the mapping coordinates (x1, y1) from the texture coordinates (u, v) to the single-path view. continuing labeling to obtain mapping coordinates (x2, y2) of a next labeling position in a single-path view; connecting the current labeling position and the next labeling position in the panoramic view to form a line, until all mapping labeling lines in each single-path view are obtained. 2.The method of claim 1, wherein, The method of acquiring a plurality of single-path views covering full circumferential direction and splicing the plurality of single-path views to obtain a panoramic view comprises the following steps: taking a shooting point as an origin, acquiring a plurality of single-path views in sequence along the circumferential direction of the shooting point, so that the fields of view of all the single-path views are superimposed to cover the full circumferential direction of the shooting point; splicing all the single-path views in the order of acquisition to obtain a panoramic view of the full circumferential direction of the shooting point.

3. The method of claim 1 or 2, wherein, The plurality of single-path views at least include a front view, a left front view, a right front view, a left side view, a right side view, a left rear view and a right rear view taking the shooting point as the origin.

4. The method of claim 1, wherein, The method of connecting the current labeling position and the next labeling position in the panoramic view to form a line, until all mapping labeling lines in each single-path view are obtained, comprises the following steps: connecting the current labeling position and the next labeling position in the panoramic view to form a line, to obtain a line segment labeled in the panoramic view; traversing all the labeling positions in the panoramic view, until all the labeling lines in the panoramic view are drawn; obtaining all the mapping labeling lines in each single-path view according to the mapping relationship between the panoramic view and each single-path view.

5. A labeling system for 360 panorama images, characterized by, The method comprises the following steps: a conversion module is configured to acquire a plurality of single-path views covering full circumferential direction, splice the plurality of single-path views to obtain a panoramic view, calculate coordinate conversion parameters between the panoramic view and each single-path view, and compile corresponding coordinate conversion formulas; the method comprises the following steps: Set panorama map coordinate x i Turn single road view coordinate x r The formula is , wherein P is a projection matrix of camera intrinsic and extrinsic parameters, M is a 3 3 coordinate conversion matrix Assume (x o , y o ) is the coordinate in the panorama, (x, y) is the coordinate in the single view, then the conversion formula from the panorama coordinate (x o , y o ) to the single view coordinate (x, y) is: x = ((M 22 M 32 y o )(M 33 x o M 13 ) (M 12 M 32 x o )(M 33 y o M 23 )) / ((M 22 M 32 y o )(M 11 M 31 x o ) (M 12 M 32 x o )(M 21 M 31 y o )) y = ((M 21 M 31 y o )(M 33 x o M 13 ) (M 11 M 31 x o )(M 33 y o M 23 )) / ((M 21 M 31 y o )(M 12 M 32 x o ) (M 11 M 31 x o )(M 22 M 32 y o )) where M 11 is 3 3 the value of the coordinate conversion matrix M in the first row and the first column, M 12 is 3 3 the value of the coordinate conversion matrix M in the first row and the second column, M 13 , M 21 , M 22 , M 23 , M 31 , M 32 , M 33 and so on; a labeling module is configured to acquire 3D click coordinates when labeling the panoramic view, convert the 3D click coordinates into texture coordinates in the panoramic view, and map the texture coordinates to each single-path view according to the coordinate conversion parameters; the method comprises the following steps: loading and displaying the 3D panoramic view, and labeling the 3D panoramic view by clicking at a current labeling position; acquiring 3D click coordinates (x', y', z') of a labeled point, and calculating texture coordinates (u, v) on the panoramic view corresponding to the 3D click coordinates (x', y', z') according to a spherical coordinate to texture coordinate conversion formula; Based on the directional differences between each individual view, coordinate transformation parameters are set between the panoramic image and each individual view, and the texture coordinates (u, v) are used as the panoramic image coordinates (x, v). o y o Substituting the coordinate transformation formula between the panoramic image and each single-path view, we obtain the mapping coordinates (x1, y1) from the texture coordinates (u, v) to the single-path view. continuing labeling to obtain mapping coordinates (x2, y2) of a next labeling position in a single-path view; connecting the current labeling position and the next labeling position in the panoramic view to form a line, until all mapping labeling lines in each single-path view are obtained.

6. An electronic device, comprising: A computer program product comprising a memory and a processor configured to implement the steps of the method of any one of claims 1-4 when executing a computer program stored in the memory.

7. A computer-readable storage medium, characterized in that, A computer program product comprising a memory and a processor configured to implement the steps of the method of any one of claims 1-4 when executing a computer program stored in the memory.

Images