Coding and decoding method, encoder, decoder and storage medium

By using the placeholder mode of parent-child nodes in the octree structure for encoding and decoding, the problem of failure to effectively utilize neighboring node information in the prior art is solved, and efficient compression and decoding of point cloud data is realized.

CN115529463BActive Publication Date: 2024-11-05BEIJING XIAOMI MOBILE SOFTWARE CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210788141.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-06-24
Publication Date
2024-11-05
Estimated Expiration
2040-06-24

AI Technical Summary

Technical Problem

The existing point cloud compression method fails to effectively utilize the information of neighboring nodes, which makes it difficult to effectively reduce the amount of data during encoding and decoding, especially in sparse point cloud data, where the compression gain is insufficient.

Method used

By determining the encoding method (planar encoding method or angle encoding method) in the octree structure, the encoding context information of the current child node, including the placeholder mode of the current parent node and the adjacent parent node, is used to generate and decode the bit stream. This provides more accurate point cloud data compression.

Benefits of technology

By considering the information of neighboring nodes, the compression amount of point cloud data is significantly reduced, and the efficiency of encoding and decoding is improved, especially in sparse point cloud data, which achieves more reliable data prediction and compression effects.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115529463B_ABST
    Figure CN115529463B_ABST
Patent Text Reader

Abstract

The present application discloses a method for encoding and decoding a point cloud, as well as an encoder and a decoder. A method for encoding a point cloud to generate a bit stream of compressed point cloud data, wherein the geometric structure of the point cloud is represented by an octree-based structure, the octree-based structure includes a plurality of nodes having a parent-child relationship by recursively dividing a volume space containing the point cloud into sub-volumes, each sub-volume is associated with a node of the octree-based structure, and the method includes the following steps: determining an encoding method, wherein the encoding method includes a plane encoding method and an angle encoding method; obtaining encoding context information of a current child node, wherein if the encoding method is a plane encoding method, the encoding context information includes plane context information, wherein the plane context information of the current child node is determined according to an occupation mode of a current parent node and an occupation mode of at least one adjacent parent node adjacent to the current parent node; if the encoding method is an angle encoding method, the encoding context information of the current child node includes the plane information of the current parent node and the occupation of a vertically adjacent parent node adjacent to the current parent node. Finally, entropy encoding is performed on the occupation of the current child node based on the determined encoding context information to generate encoding data for a bit stream.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] Divisional Application Statement

[0002] This application is a divisional application based on Chinese invention patent application number 202080003878.4, application date June 24, 2020, and invention name “Encoding and decoding methods, encoder, decoder and storage medium”. Technical Field

[0003] The present application generally relates to point cloud compression. In particular, the present application relates to encoding and decoding methods and encoders and decoders for improving entropy coding of point clouds. Background Art

[0004] 3D point clouds have recently become a popular representation of 3D media information, replacing 3D meshes. The use cases associated with point cloud data are very diverse, including:

[0005] 3D assets in film production,

[0006] 3D assets for real-time 3D immersive telepresence or virtual reality (VR) applications,

[0007] 3D free viewpoint video (e.g. for watching sports),

[0008] Geographic Information Systems (Cartography),

[0009] Cultural heritage (storing fragile assets in digital form),

[0010] Autonomous driving (large 3D environment mapping), etc.

[0011] A point cloud is a collection of points in 3D space, each with associated properties such as color, material properties, etc. Point clouds can be used to reconstruct objects or scenes as a combination of these points. Point clouds can be acquired using multiple cameras and depth sensors with different settings, and can consist of thousands or even billions of points in order to realistically reproduce the reconstructed scene.

[0012] The location of each point of the point cloud needs to be stored (usually encoding X, Y, Z information as 32-bit or 64-bit floating point numbers) and its attributes (usually at least 24-bit encoded RGB colors). Sometimes there are billions of points in the point cloud, so it is easy to explain that the raw data of the point cloud can be as large as several gigabytes. Therefore, compression techniques are urgently needed to reduce the amount of data required to represent the point cloud.

[0013] Two different approaches have been developed for point cloud compression:

[0014] First, in the video-based point cloud compression (VPCC) method, point cloud compression is performed in the following way: the point cloud is projected multiple times on 3 different axes X, Y, Z and different depths so that all points appear in one projected image. Then, the multiple projected images obtained by projection are processed into multiple parts (patches) to eliminate redundancy and rearranged to obtain the final image, in which the attached metadata is used to convert the pixel position into the point position in space. Then, compression is performed using a traditional image / video MPEG encoder. The advantage of this method is that it reuses existing encoders and naturally supports dynamic point clouds (using video encoders), but this method is hardly applicable to sparse point clouds, and it is expected that the compression gain can be improved using point cloud-specific methods.

[0015] Secondly, in the geometry-based point cloud compression (GPCC) method, point positions (often referred to as geometry) and point attributes (color, transparency, etc.) are encoded separately. In order to encode the geometry, an octree structure is used. The entire point cloud is fitted into a cube, and the cube is continuously divided into eight sub-cubes until each sub-cube contains only a single point. Therefore, the position of the point is replaced by the placeholder information of each node of the tree. Since each cube has only 8 sub-cubes, 3 bits are sufficient to encode the placeholder, so for a tree of depth D, 3 bits are required. D bits are used to encode the position of a point. While this transformation alone is not sufficient to provide significant compression gains, it should be noted that due to the tree structure, many points share the same node value, and due to the use of an entropy encoder, the amount of information can be significantly reduced.

[0016] Knowing that many point clouds include surfaces, in the current design of GPCC, a plane encoding scheme is introduced to more efficiently encode such nodes suitable for octrees.

[0017] Therefore, a flag isPlanar is introduced, which indicates whether the placeholder child nodes belong to the same horizontal plane. If isPlanar is true, an additional bit is used to mark planePosition to indicate whether the plane is the lower plane or the upper plane.

[0018] The isPlanar flag is encoded by using a binary arithmetic encoder, where 8 (i.e., 2×2×2) bits of context information are used as plane context information. In addition, if the node is planar, the plane position information is encoded by using a binary arithmetic encoder, where 24 (i.e., 2×3×2×2) bits of context information are used as plane position context information.

[0019] In addition, an angle encoding scheme is introduced to enhance the plane encoding scheme with the angle context for sparse point cloud data, where neighboring occupancy tends to be zero (i.e., empty) in most cases. More specifically, when planePosition is encoded with reduced context information, the angle encoding scheme replaces the context information, and the isPlanar flag is still arithmetically encoded using the same context model as in the plane encoding scheme.

[0020] However, in current methods for encoding and decoding point clouds, the information of neighboring nodes is not considered.

[0021] Therefore, the object of the present invention is to provide a coding and decoding method as well as an encoder and a decoder to enable improved fast and accurate data compression of point clouds. Summary of the invention

[0022] In one aspect of the present invention, a method for encoding a point cloud to generate a bitstream of compressed point cloud data is provided, wherein the geometric structure of the point cloud is represented by an octree-based structure, the octree-based structure comprising a plurality of nodes having a parent-child relationship by recursively partitioning a volume space containing the point cloud into sub-volumes, each of the sub-volumes being associated with a node of the octree-based structure, the method comprising the following steps:

[0023] Determine a coding method, wherein the coding method includes a plane coding method and an angle coding method;

[0024] Get the encoding context information of the current child node; where,

[0025] If the encoding mode is a plane encoding mode, the encoding context information includes plane context information, wherein the plane context information of the current child node is determined according to the placeholder mode of the current parent node and the placeholder mode of at least one adjacent parent node immediately adjacent to the current parent node;

[0026] If the encoding method is an angle encoding method, the encoding context information of the current child node includes the plane information of the current parent node and the placeholders of the vertically adjacent parent nodes adjacent to the current parent node; and

[0027] The placeholder of the current child node is entropy encoded based on the determined encoding context information to generate encoding data for a bitstream.

[0028] Therefore, a coding method is determined first, and context information is determined according to the coding method. The context information includes two different context information, so as to perform effective entropy coding.

[0029] Wherein, for the plane coding method, the current parent node and the adjacent parent node are at the same depth D of the tree. The current parent node and the adjacent parent node share a common surface due to their positions. For the current child node, which is a direct child node of the current parent node at a depth D+1 of the tree, the plane context information is determined depending on the occupancy pattern of the current parent node and at least one adjacent parent node. Wherein, the occupancy pattern includes the distribution of the placeholder child nodes within the corresponding parent node. Wherein, the plane context information includes information about possible planes or surfaces represented by the point cloud passing through the current child node. Therefore, information about the occupancy pattern of at least one adjacent node of the current parent node is considered as context information for entropy coding.

[0030] For the angle encoding approach, a more efficient plane context approach is implemented instead. Compared with the plane context information of the plane encoding approach, only the information accessible in the sparsely distributed cloud data is considered.

[0031] As a result, the occupancy of the current child node can be predicted more reliably, which can be used by the entropy encoder to further reduce the amount of data. In this way, the entire tree is traversed to determine the occupancy of each node and provide sufficient context information for the entropy encoder.

[0032] Preferably, the adjacent parent node is a node that is immediately adjacent to the current child node and shares a surface with the current child node.

[0033] Preferably, for the plane coding mode, the coding context information includes plane context information, wherein unique plane context information of the current child node is determined according to the occupancy mode of the current parent node and the occupancy mode of at least one adjacent parent node immediately adjacent to the current parent node.

[0034] Preferably, for the angle coding method, the coding context information of the current child node includes the plane information of the current parent node and the placeholders of the vertically adjacent parent nodes adjacent to the current parent node.

[0035] Preferably, if the encoding mode is a plane encoding mode, the encoding context information includes plane position context information, wherein the plane position context information includes one or more of the following items:

[0036] The distance d to the nearest encoded node at the same coordinates in the corresponding parent node at the same depth D in the octree-based structure;

[0037] The planar position of the most recently encoded node at the same coordinates in the corresponding parent node at the same depth D in the octree-based structure;

[0038] The plane position of the current child node is determined according to the placeholder mode of at least one adjacent parent node immediately adjacent to the current parent node.

[0039] Therein, in particular, the distance d is discretized into the values ​​"near", "not too far" and "far". If there are encoded nodes at the same coordinates at the same depth D, the plane position, i.e. "high" or "low", can also be included in the plane position context information. Therein, the coordinates of the child node are the relative position within the corresponding parent node. In addition, the plane position may be included in the plane position context information determined based on the occupancy pattern of at least one adjacent parent node adjacent to the current parent node. In particular, the plane position context information includes more than one or all of the above information. In particular, in the plane coding method, the plane position context information may include the above information.

[0040] Preferably, if there is no most recently encoded node at the same coordinates in the corresponding parent node at the same depth D in the octree-based structure, the plane position context information only includes the plane position of the current child node determined based on the occupancy pattern of at least one adjacent parent node adjacent to the current parent node. Therefore, in this case, the plane position context information is determined only based on information collected from adjacent nodes. Further information is not available and therefore cannot be used. However, since information from adjacent nodes is included, effective compression can be achieved even in this case. In particular, the axisIdx statement proposed in the current GPCC specification to identify the axis perpendicular to the plane used to encode planePosition is replaced by context information determined based on the occupancy of adjacent nodes.

[0041] Preferably, if the encoding mode is an angle encoding mode and the current child node is not applicable to the angle encoding mode, the encoding context information includes plane position context information, wherein the plane position context information includes one or more of the following items:

[0042] The distance d to the nearest encoded node at the same coordinates in the corresponding parent node at the same depth D in the octree-based structure;

[0043] The planar position of the most recently encoded node at the same coordinates in the corresponding parent node at the same depth d in the octree-based structure;

[0044] The plane position of the current child node according to the placeholder pattern of at least one adjacent parent node immediately adjacent to the current parent node.

[0045] Therefore, in this case, the plane position context information may be the same as the plane position information in the plane coding mode. In particular, the plane position context information includes one or more or all of the above information. In particular, in the angle coding mode, if the current child node is not applicable to the angle coding mode, the plane position context information may include the above information.

[0046] Preferably, in order to determine the plane context information and the plane position context information according to the occupancy mode of at least one adjacent node, the current plane position of the current child node relative to the current parent node is determined, wherein the current plane position includes a high plane position and a low plane position. Wherein, for a certain geometric axis under consideration, along the axis passing through the parent node, if there is first a plane perpendicular to the axis in the current parent node and including the child nodes of the current child node, the current plane position is a low plane position. If there is first a plane perpendicular to the axis in the current parent node and not including the child nodes of the current child node, the current plane position is a high plane position. Therefore, the current plane position is determined for the plane context information under the plane coding mode, and a part of the plane position context information under the plane coding mode and the angle coding mode.

[0047] Preferably, if the current plane position is a low plane position, the plane context information and the plane position context information are determined based on: the occupancy of a first group of four child nodes of an adjacent parent node that are adjacent to the current parent node and the occupancy of a second group of four child nodes of the current parent node that are directly opposite to the first group. In other words, the first group of four child nodes of the adjacent parent node are adjacent to a common surface of the current parent node and the adjacent parent node, and the second group is arranged on the opposite side of the current parent node adjacent to the current parent node. Each group is preferably defined by four child nodes to form a plane, wherein each plane is perpendicular to the geometric axis under consideration and is adjacent to the current child node to be evaluated. Therefore, sufficient information can be obtained to provide a reliable prediction of the likelihood of a plane in the current child node and the likelihood of its position.

[0048] Preferably, for the current plane position being in a low plane position, if at least one child node of the first group is occupied and the second group is not occupied, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of the low plane position.

[0049] Preferably, for the current plane position being in a low plane position, if the first group is not occupied and at least one child node of the second group is occupied, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of a high plane position.

[0050] Preferably, for the current plane position being in a low plane position, if the first group is not occupied and the second group is not occupied, the plane context information includes the possibility of the plane, wherein if at least one or more child nodes in the adjacent parent node that do not belong to the first group are occupied, the plane position context information includes the possibility of the low plane, otherwise it is "UNKNOWN".

[0051] Preferably, for the current plane position being in a low plane position, if at least one child node of the first group is occupied and at least one child node of the second group is occupied, and #FirstGroup>#SecondGroup+1, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of the low plane position, wherein #FirstGroup indicates the number of placeholder nodes in the first group and #SecondGroup indicates the number of placeholder nodes in the second group.

[0052] Preferably, for the current plane position being in a low plane position, if at least one child node of the first group is occupied and at least one child node of the second group is occupied, and #FirstGroup<#SecondGroup+1, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of a high plane position; wherein #FirstGroup indicates the number of placeholder nodes in the first group, and #SecondGroup indicates the number of placeholder nodes in the second group.

[0053] Preferably, if the current plane position is high, the plane context information and the plane position context information are determined according to the following: the occupancy pattern of the adjacent parent node immediately behind the current parent node along the axis under consideration, and the occupancy of the group of four child nodes of the current parent node in the plane perpendicular to the axis under consideration and adjacent to the current child node. In other words, the occupancy of the child nodes of the adjacent parent node adjacent to the common surface of the current parent node (i.e., after the common surface) is unknown due to the decoding order. Therefore, the occupancy of the adjacent parent node itself that shares a common surface with the current child node is used together with the occupancy of the following group: the group is arranged on the opposite side of the current child node in the current parent node, i.e., opposite to the adjacent parent node (i.e., in the encoding / decoding order). Therefore, the child node group does not include the current child node to be encoded. Among them, each group is preferably defined by four child nodes. A plane, wherein each plane is perpendicular to the geometric axis under consideration and adjacent to the current child node to be evaluated in the current parent node. Therefore, sufficient information can be obtained to provide a reliable prediction of the possibility of the plane in the current child node and the possibility of its position.

[0054] Preferably, for the current plane position being in a high plane position, if at least one child node of the group is occupied and an adjacent parent node is not occupied, the plane context information includes the likelihood of a plane, and the plane position context information includes the likelihood of a low plane position.

[0055] Preferably, for the current plane position being in a high plane position, if the group is not occupied and the adjacent parent node is occupied, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of the high plane position.

[0056] Preferably, for the current plane position being in a high plane position, the plane context information includes the possibility of the plane if the group is not occupied and the adjacent parent node is not occupied, wherein the plane position context information includes the possibility of a low plane if at least one or more child nodes in the adjacent parent node immediately preceding the current parent node along the axis considered are occupied, otherwise it is "UNKNOWN".

[0057] Preferably, for the current plane position being in a high plane position, if at least one child node of the group is occupied and the adjacent parent node is occupied, and #Group>2, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of the low plane position, where #Group indicates the number of placeholder nodes in the group. It is assumed that statistically one child node in the adjacent parent node is occupied.

[0058] Preferably, for the current plane position being in a high plane position, if at least one child node of the group is occupied and the adjacent parent node is occupied, and #Group<2, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of a high plane position, wherein #Group indicates the number of occupied nodes in the group.

[0059] Preferably, for the current plane position being at a high plane position, if at least one child node of the group is occupied and the adjacent parent node is occupied, and #Group=2, the plane context information includes the possibility that there is no plane and the plane position is inferred to be unknown; wherein #Group indicates the number of placeholder nodes in the group.

[0060] Preferably, the neighboring parent nodes are arranged along one of the geometric axes. If the plane context information is evaluated along one of the geometric axes (X, Y, Z), the neighboring parent node is located immediately before or after the current parent node along the specific axis.

[0061] Preferably, the plane context information is determined along all three geometric axes. Thus, the plane or surface represented by the points of the point cloud is evaluated along each axis.

[0062] Preferably, the current parent node is arranged relative to the adjacent parent nodes in the order of decoding. During decoding, the nodes of the octree-based structure are evaluated successively. Therefore, information about the placeholder pattern before the current parent node in the decoding order is already accessible and can therefore be used to determine the plane context information for decoding. In particular, the decoding order is completely opposite to the encoding order. However, during encoding, the exact position of each point in the point cloud is known.

[0063] In one aspect of the present invention, a method for decoding a bit stream of compressed point cloud data to generate a reconstructed point cloud, wherein the geometric structure of the point cloud is represented by an octree-based structure, the octree-based structure comprising a plurality of nodes having a parent-child relationship by recursively partitioning a volume space containing the point cloud into sub-volumes, each of the sub-volumes being associated with a node of the octree-based structure, comprises the following steps:

[0064] Determine a coding mode, wherein the coding mode includes a plane coding mode and an angle coding mode;

[0065] Get the encoding context information of the current child node, where

[0066] If the encoding mode is the plane encoding mode, the encoding context information includes plane context information, wherein the plane context information of the current child node is determined according to the placeholder mode of the current parent node and the placeholder mode of at least one adjacent parent node immediately adjacent to the current parent node;

[0067] If the encoding mode is the angle encoding mode, the encoding context information of the current child node includes the plane information of the current parent node and the placeholder of the vertically adjacent parent node adjacent to the current parent node; and

[0068] The bitstream is entropy decoded based on the encoding context information of the current child node to reconstruct the point cloud.

[0069] Wherein, for the plane coding method, the current parent node and the adjacent parent node are at the same depth D of the tree. The current parent node and the adjacent parent node share a common surface due to their positions. For the current child node, which is a direct child node of the current parent node at a depth D+1 of the tree, the plane context information is determined depending on the occupancy mode of the current parent node and at least one adjacent parent node. Wherein, the plane context information includes information about possible planes or surfaces represented by a point cloud passing through the current child node. Therefore, information about the occupancy of at least one adjacent node of the current parent node is considered as context information for entropy decoding.

[0070] For the angle encoding approach, a more efficient plane context approach is implemented instead. Compared with the plane context information of the plane encoding approach, only the information accessible in the sparsely distributed cloud data is considered.

[0071] As a result, the occupancy of the current child node can be predicted more reliably, which can be used by the entropy decoder to further reduce the amount of data. In this way, the entire tree is traversed to determine the occupancy of each node and provide sufficient context information for the entropy encoder.

[0072] Preferably, the decoding method is further constructed according to the above features about the encoding method. These features can be freely combined with the decoding method.

[0073] In one aspect of the present invention, an encoder is provided for encoding a point cloud to generate a bit stream of compressed point cloud data, wherein the geometric structure of the point cloud is represented by an octree-based structure, the octree-based structure comprising a plurality of nodes having a parent-child relationship by recursively partitioning a volume space containing the point cloud into sub-volumes, each of the sub-volumes being associated with a node of the octree-based structure, the encoder comprising:

[0074] processor, and

[0075] A memory storage device, wherein the memory storage device stores instructions executable by the processor, and when the instructions are executed, the processor is prompted to perform the above-mentioned method for encoding and decoding.

[0076] In one aspect of the present invention, there is provided a decoder for decoding a bit stream of compressed point cloud data to generate a reconstructed point cloud, wherein the geometric structure of the point cloud is represented by an octree-based structure, the octree-based structure comprising a plurality of nodes having a parent-child relationship by recursively partitioning a volume space containing the point cloud into sub-volumes, each of the sub-volumes being associated with a node of the octree-based structure, the decoder comprising:

[0077] processor, and

[0078] A memory storage device, wherein the memory storage device stores instructions executable by the processor, and when the instructions are executed, the processor is prompted to perform the above decoding method.

[0079] In one aspect of the present invention, a non-transitory computer-readable storage medium is provided, which stores instructions executed by a processor. When the instructions are executed by the processor, the processor is prompted to perform the above encoding and / or decoding method. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] Reference will now be made by way of example to the accompanying drawings, in which exemplary embodiments of the present application are shown, in which:

[0081] Figure 1 is a block diagram showing an overall diagram of a point cloud encoder,

[0082] Figure 2 is a block diagram showing an overall diagram of a point cloud decoder,

[0083] Figure 3 This is a schematic diagram of the octree data structure.

[0084] Figure 4 The eight child nodes in each node are numbered.

[0085] Figure 5 shows the neighboring nodes of a node with known placeholder information,

[0086] Figure 6 is a graphical representation of contextual information,

[0087] Figure 7 A specific embodiment of the present invention is shown.

[0088] Figure 8 A specific embodiment of the present invention is shown.

[0089] Fig. 9 A specific embodiment of the present invention is shown.

[0090] Fig.10 is a flowchart showing the steps of the first embodiment and the second embodiment of the present invention,

[0091] Fig.11 is a schematic diagram of the encoder device, and

[0092] Fig.12 is a schematic diagram of a decoder device. DETAILED DESCRIPTION

[0093] The present application describes a method for encoding and decoding a point cloud and an encoder and decoder for encoding and decoding a point cloud. A current parent node associated with a child is further divided into a plurality of children, each further child corresponding to a child of the current parent node, and the encoder determines the occupancy mode of the current parent node based on the occupancy state of the child node. Planar context information is determined based on the occupancy mode of the current parent node and the occupancy mode of the adjacent parent node immediately adjacent to the current parent node. The entropy encoder encodes the occupancy mode based on the determined plane context information to generate encoded data for a bitstream. The decoder determines the same plane context information and entropy decodes the bitstream to reconstruct the occupancy mode.

[0094] Other aspects and features of the present application will be apparent to those skilled in the art by reading the following description of the examples in conjunction with the accompanying drawings.

[0095] Sometimes in the following description, the terms "node" and "sub-volume" may be used interchangeably. It should be understood that a node is associated with a sub-volume. A node is a specific point on a tree, which may be an internal node or a leaf node. A sub-volume is a bounded physical space represented by a node. The term "volume" may be used to refer to the largest bounded space defined to contain a point cloud. In order to construct a tree structure of interconnected nodes to encode point cloud data, a "volume" is recursively divided into sub-volumes.

[0096] A point cloud is a collection of points in a three-dimensional coordinate system. These points are generally intended to characterize the outer surface of one or more objects. Each point has a location (position) in a three-dimensional coordinate system. The position can be represented by three coordinates (X, Y, Z) in a Cartesian coordinate system or any other coordinate system. These points may have other associated attributes such as color, which in some cases may be three-component values ​​such as R, G, B or Y, Cb, Cr. Other associated attributes may also include transparency, reflectivity, normal vector, etc., depending on the desired application of the point cloud data.

[0097] Point clouds can be static or dynamic. For example, a detailed scan or map of an object or terrain can be static point cloud data. LiDAR-based environmental scans for machine vision purposes can be dynamic point cloud data because the point cloud (at least potentially) changes over time (e.g., with each successive scan of the "volume"). Thus, a dynamic point cloud is a sequence of point clouds ordered in time.

[0098] Point cloud data can be used in many applications, including, for example, preservation (scanning of historical artifacts), mapping, machine vision (e.g., autonomous or semi-autonomous vehicles), and virtual reality systems. Dynamic point cloud data for applications such as machine vision can be quite different from static point cloud data for purposes such as preservation. For example, automotive vision typically involves relatively low resolution, colorless, highly dynamic point clouds acquired by LiDAR (or similar) sensors with high acquisition frequencies. Such point clouds are not intended for human use or viewing, but rather for machine object detection / classification in decision-making processes. For example, a typical LiDAR frame contains tens of thousands of points in sequence, while high-quality virtual reality applications require millions of points. It can be expected that higher resolution data will be needed over time as computing speeds increase and new applications emerge.

[0099] Despite the usefulness of point cloud data, the lack of effective and efficient compression, i.e. encoding and decoding processing, can hinder adoption and deployment.

[0100] One of the more common mechanisms for encoding point cloud data is to use a tree-based structure. In a tree-based structure, a bounded three-dimensional volume of a point cloud is recursively divided into sub-volumes. The nodes of the tree correspond to sub-volumes. Whether to further divide the sub-volume can be determined based on the resolution of the tree and / or whether the sub-volume contains any points. A leaf node can have a placeholder flag that indicates whether its associated sub-volume contains points. A split flag can indicate whether a node has child nodes (i.e., whether the current volume has been further split into sub-volumes). In some cases, these flags can be entropy encoded, and in some cases, predictive coding can be used.

[0101] A commonly used tree structure is the octree. In this structure, each volume / subvolume is a cube, and each division of a subvolume further produces eight subvolumes / subcubes. An example of this tree structure is Figure 3 , which has a node 30 that can represent a volume containing a complete point cloud. The volume is divided into eight sub-volumes 32, each sub-volume and Figure 3 A node containing a point indicates a placeholder node 34 containing at least one point 35 of the point cloud, while an empty node 36 represents a sub-volume that does not contain a point of the point cloud. Figure 3 As shown, the placeholder node can be further divided into eight child fonts, and these eight fonts are respectively associated with each child node 38 of a specific parent node 40 to determine the placeholder mode of the parent node 40. Figure 3 As shown, the placeholder pattern of the exemplary parent node 40 can be represented in binary form as "00100000", which indicates the occupied third child node 38. In some embodiments, such a placeholder pattern is encoded by a binary entropy encoder to generate a bit stream of point cloud data.

[0102] Reference now Figure 1 , which shows a simplified block diagram of a point cloud encoder 10 according to aspects of the present application. The point cloud encoder 10 receives point cloud data and may include a tree construction module for generating an octree that represents the geometry of a volume space containing a point cloud and indicates the location or position of points of the point cloud in the geometry.

[0103] The basic processing for creating an octree to encode a point cloud can include:

[0104] 1. Start with a bounded volume (cube) containing the point cloud in a coordinate system;

[0105] 2. Split the bounded volume into 8 sub-volumes (eight sub-cubes);

[0106] 3. For each sub-volume, if the sub-volume is empty, mark the sub-volume with 0; or if there is at least one point in the sub-volume, mark the sub-volume with 1;

[0107] 4. For all sub-volumes marked as 1, repeat (2) to split these sub-volumes until the maximum split depth is reached; and

[0108] 5. For all leaf cubes (subcubes) of maximum depth, mark it as 1 if the leaf cube is not empty, otherwise mark it as 0.

[0109] The tree may be traversed in a predefined order (within each partitioned sub-volume, either breadth-first or depth-first, and according to a scan mode / order) to generate a bit sequence representing the placeholder pattern of each node.

[0110] The bit sequence may then be encoded using an entropy encoder 16 to produce a compressed bitstream 14. The entropy encoder 16 may encode the bit sequence using a context model 18, which specifies the probability of the entropy encoder 16 encoding the bit based on the context determination. The context model 18 may be adaptively updated after encoding each bit or a defined set of bits.

[0111] Like video coding or image coding, point cloud coding can include predictive operations, where an effort is made to predict the pattern of a subvolume and the residual of the prediction is encoded instead of the pattern itself. The prediction can be spatial (depending on previously encoded subvolumes in the same point cloud) or temporal (depending on previously encoded point clouds in a time-ordered sequence of point clouds).

[0112] exist Figure 2 A block diagram of an exemplary point cloud decoder 20 corresponding to the encoder 10 is shown in FIG. The point cloud decoder 20 includes an entropy decoder 22 that uses the same context model 24 as used by the encoder 10. The entropy decoder 22 receives an input bit stream 26 of compressed data and entropy decodes the data to produce an output sequence of decompressed bits. This sequence is then converted to reconstructed point cloud data by a tree reconstructor. The tree reconstructor reconstructs a tree structure 28 based on the decompressed data and knowledge of the scan order in which the tree data is binarized. Thus, the tree reconstructor is able to reconstruct the positions of points in the point cloud.

[0113] Reference Figure 4 , which shows that the parent node 112 is divided into eight child nodes 110, and the eight child nodes 110 are 2×2×2 cubes, where each cube has the same size and the side length is half the side length of the cube associated with the parent node 112. In addition, Figure 4 Indicates the number used by the child node 110 within the parent node 112. In further description, Figure 4 The numbering system shown in . Figure 4The spatial orientation of the parent node 112 shown in three-dimensional space represented by geometric axes X, Y, Z is also indicated.

[0114] Among them, the occupancy pattern may include plane information about the probability of whether a particular node is occupied because the points in the node belong to a surface. Typically, in the real world, closed surfaces dominate. This is especially true for indoor rooms, but also for urban outdoor scenes. The entropy encoder and decoder take advantage of this fact. If a surface represented by a point cloud can be detected, the distribution of points on this surface can be predicted, so that the occupancy probability of a particular node belonging to the surface can be determined. This can be done by defining context information, which is used to encode and decode the bitstream using the isPlanar flag. Among them, the plane context information is usually a binary value, where the set isPlanar flag (isPlanar=1) of a particular node is interpreted as the node is likely to belong to a particular surface. In addition to the fact that there is a surface in the node, other plane information can also be considered, for example, the plane position information implemented by the planePosition flag, which indicates the position of the plane within the current child node. The planePosition flag can also be a binary value with values ​​"high" and "low" indicating the corresponding position. The entropy encoder / decoder encodes the plane information into the bitstream by using the plane context information, thereby reducing the data of the bitstream.

[0115] In addition, there are other encoding methods that use different context information for different types of point clouds. Therefore, a plane encoding method and an angle encoding method are defined. Among them, the angle encoding method is preferably applied to sparsely distributed point clouds, with the intention of providing angle context information to the entropy encoder / decoder. Among them, the encoding method can be indicated in the metadata of the bitstream.

[0116] Therefore, according to the present disclosure, in the case of a plane coding method, the context information includes the plane context information and the plane position context information as described above. The plane context information is determined according to the placeholder mode of the current parent node of the current child node to be encoded and the placeholder mode of at least one of the neighboring nodes, as described in more detail below. In addition, Figure 6 Describes the plane location context information. The plane location context information includes the following information:

[0117] a) the distance d to the nearest encoded node 121 at the same coordinates at depth D in the octree structure; d is discretized into three values: “near” or “not too far” or “far”;

[0118] b) the planar position of the nearest encoded node 121 at the same depth (low or high) and at the same coordinates (if any);

[0119] c) The planar position of the current child node according to the placeholder pattern of at least one adjacent parent node, as described in more detail below.

[0120] Among them, the plane context information is 1 bit (plane or no plane), and the plane position context information is 3×2×2 bits of information.

[0121] In addition, according to the present disclosure, in the case of angle coding, the context information includes the plane context information and plane position context information as described above. The plane context information includes:

[0122] a) the isPlanar flag of the current parent node, that is, whether there is a plane in the current parent node; and

[0123] b) The placeholder of the vertically adjacent parent node 124 that is vertically adjacent to the current parent node.

[0124] In addition, the plane position context information used for the angle encoding method includes the following information:

[0125] a) The planePosition of an available node identified by the angle between the available node and the current child node.

[0126] Reference now Figure 5 , which shows a current parent node 100 having two child nodes 102 and 103, the two child nodes 102 and 103 being illustrated as placeholder nodes, i.e., the sub-volumes of the two child nodes each contain at least one point of the point cloud. Figure 5Indicated in are adjacent parent nodes 104, 106, 108, each of which itself includes at least one placeholder child node. Wherein, before predicting the plane context information of the child nodes 102, 103 located at a specific octree level, the placeholder information of the three adjacent adjacent parent nodes 104, 106, 108 at the same level D as the current parent node 100 in the octree is known on the decoder side. Due to the given decoding order, when evaluating the child nodes 102, 103 of the current parent node 100, the information about the adjacent parent nodes 104, 108, 106 is known on the decoder side and can be used to determine the context information. Since these three adjacent parent nodes 104, 106, 108 are adjacent to the current child nodes 102, 103, there is a correlation between the placeholder patterns of these adjacent parent nodes 104, 106, 108 and the placeholder patterns of the current child nodes 102, 103 to be encoded. Wherein, the placeholder pattern refers to the placeholders of the child nodes of each parent node. This information is readily available but is not currently used in existing encoding and decoding methods. For example, if the context information of the current child node 102 is evaluated along the X-axis, the adjacent parent node 104 that is immediately adjacent to the current parent node 100 along the X-axis can be taken into account. In addition, as described above, the immediately adjacent parent node is considered, which is before the current parent node in the order of decoding. Therefore, during decoding, the placeholder pattern of the previous node is known to the decoder and can be used to determine the context information. If the context information is evaluated along the Y-axis, the adjacent parent node 106 that is adjacent to the current parent node 100 along the Y-axis can be taken into account. This information is also known to the decoder and can be easily used to obtain reliable plane context information for the current child node 102.

[0127] However, when considering the correlation between the occupancy of the current child node and the occupancy pattern of at least one adjacent parent node, the position of the current child node to be encoded must also be considered. In this way, the current plane position of the current child node is to be determined. Therein, the current plane position indicates the plane of the current child node within the current parent node along the axis considered and the plane perpendicular to the geometric axis. For example, along the X-axis, the first low plane is defined by child nodes 0, 1, 2, 3, and the high plane is defined by child nodes 4, 5, 6, 7. Along the Y-axis, the low plane is defined by child nodes 0, 1, 4, 5, and the high plane is defined by child nodes 2, 3, 6, 7. Along the Z-axis, the low plane is defined by child nodes 0, 2, 4, 6, and the high plane is defined by child nodes 1, 3, 5, 7. Therefore, in Figure 5 In the example of , the current plane position of the child node 102 along the X-axis is the low plane position, and the current plane position of the child node 103 along the X-axis is the high plane position.

[0128] Reference now Figure 7, which shows how to determine the configuration of plane context information and plane position context information based on the occupancy of the neighboring nodes of the current child node, which can then be used in entropy coding to compress point cloud data. In this regard, one of the objectives of the present invention is to take into account the occupancy pattern of the neighboring parent nodes of the current child node 120 and incorporate such information into the context information when encoding the plane information. Figure 7 It shows the case where the current child node to be encoded is at a lower position of the current parent node.

[0129] Figure 7 (a) shows the determination of plane context information along the X-axis. The plane context information of the current child node 120 is to be evaluated. Therefore, the occupancy pattern of the current parent node 122 and the adjacent parent node 124 adjacent to the current parent node 122 is evaluated. Among them, the adjacent parent nodes 124 to be evaluated are arranged along the X-axis. In addition, the decoding order of the adjacent parent nodes 124 is arranged before the current parent node 122. Therefore, when the current parent node 122 is decoded, all information about the adjacent parent node 124 is already available. In addition, since the position and occupancy pattern of the point cloud are completely known to the encoder, the occupancy information about the adjacent parent node 124 is also known during the encoding of the information of the current child node 120.

[0130] In particular, in order to determine the context information of the current child node 120, a first group of child nodes 126 as child nodes of the adjacent parent node 124 and a second group of child nodes 128 as child nodes of the current parent node 122 are considered. Among them, the four child nodes of the first group 126 of the adjacent parent node 124 are directly adjacent to the current parent node 122. Therefore, the first group of child nodes 126 each share a common surface with the current parent node 122. The four child nodes of the second group 128 of the current parent node 122 are directly opposite to the first group 126. Therefore, the four child nodes of the second group 128 form a side of the current parent node 122, which is directly opposite to the side that is the common surface between the adjacent parent node 124 and the current parent node 122. In other words, the planes defined by the first group of child nodes 126 and the second group of child nodes 128 are located on both sides of the current child node 120 and are perpendicular to the axis under consideration. Therefore, through these two planes, a reliable determination of the context information can be provided. A surface in the point cloud that passes through the current child node 120 is also very likely to pass through one of the planes defined by the first group of child nodes 126 or the second group of child nodes 128, so the context information of the current child node 120 depends largely on the occupancy of the first group of child nodes 126 and / or the occupancy of the second group of child nodes 128.

[0131] In general, for the case where the current plane position is low, if at least one child node of the first group is occupied and the second group is not occupied, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of a low plane position. If the first group is not occupied and at least one child node of the second group is occupied, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of a high plane position. If the first group is not occupied and the second group is not occupied, the plane context information includes the possibility of the plane, wherein if at least one or more child nodes of the adjacent parent node that do not belong to the first group are occupied, the plane position context information includes the possibility of a low plane, otherwise it is "UNKNOWN". If at least one child node of the first group is occupied and at least one child node of the second group is occupied, and #FirstGroup>#SecondGroup+1, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of a low plane position. If at least one child node of the first group is occupied and at least one child node of the second group is occupied, and #FirstGroup<#SecondGroup+1, the plane context information includes the possibility of a plane, and the plane position context information includes the possibility of a high plane position, where #FirstGroup indicates the number of placeholder nodes in the first group and #SecondGroup indicates the number of placeholder nodes in the second group.

[0132] More specifically, for the evaluation along the X-axis and Figure 4 Numbering system shown:

[0133] If at least one of the child nodes 4 to 7 of the first group is occupied and the child nodes 4 to 7 of the second group are not occupied, the plane context information includes a likelihood of a plane and the plane position context information includes a likelihood of a low plane position.

[0134] If child nodes 4 to 7 of the first group are unoccupied and at least one of the child nodes 4 to 7 of the second group is occupied, the plane context information includes a likelihood of a plane and the plane position context information includes a likelihood of a high plane position.

[0135] ·If child nodes 4 to 7 of the first group are not occupied and child nodes 4 to 7 of the second group are not occupied, the plane context information includes the possibility of the plane, wherein the plane position context information includes a low possibility of the plane if at least one or more child nodes 0 to 3 of the adjacent parent node are occupied, otherwise it is "UNKNOWN".

[0136] If at least one of the child nodes 4 to 7 of the first group is occupied and at least one of the child nodes 4 to 7 of the second group is occupied, and #FirstGroup>#SecondGroup+1, the plane context information includes the possibility of a plane, and the plane position context information includes the possibility of a low plane position.

[0137] If at least one of the child nodes 4 to 7 of the first group is occupied and at least one of the child nodes 4 to 7 of the second group is occupied, and #FirstGroup<#SecondGroup+1, the plane context information includes the likelihood of a plane, and the plane position context information includes the likelihood of a high plane position.

[0138] Therefore, for encoding, when entropy encoding is performed on the current plane position which is a low position, the probability of the presence or absence of a plane perpendicular to the X-axis in the current child node is regarded as context information. Among them, this information can be used as plane context information in the plane encoding method and the angle encoding method, and can be included as plane position context information in the plane encoding method.

[0139] In order to Figure 7 (b) Evaluate along the Y axis to determine the plane context information of the current child node 120, which is set as follows:

[0140] If at least one of the child nodes 2, 3, 6, 7 of the first group is occupied and the child nodes 2, 3, 6, 7 of the second group are not occupied, the plane context information comprises a likelihood of a plane and the plane position context information comprises a likelihood of a low plane position.

[0141] If the child nodes 2, 3, 6, 7 of the first group are not occupied and at least one of the child nodes 2, 3, 6, 7 of the second group is occupied, the plane context information comprises a likelihood of a plane and the plane position context information comprises a likelihood of a high plane position.

[0142] ·If the child nodes 2, 3, 6, 7 of the first group are not occupied and the child nodes 2, 3, 6, 7 of the second group are not occupied, the plane context information includes the possibility of the plane, wherein if at least one or more of the child nodes 0, 1, 4, 5 of the adjacent parent node are occupied, the plane position context information includes the possibility of a low plane, otherwise it is "UNKNOWN".

[0143] If at least one of the child nodes 2, 3, 6, 7 of the first group is occupied and at least one of the child nodes 2, 3, 6, 7 of the second group is occupied, and #FirstGroup>#SecondGroup+1, the plane context information includes the possibility of a plane, and the plane position context information includes the possibility of a low plane position.

[0144] If at least one of the child nodes 2, 3, 6, 7 of the first group is occupied and at least one of the child nodes 2, 3, 6, 7 of the second group is occupied, and #FirstGroup<#SecondGroup+1, the plane context information includes the possibility of a plane, and the plane position context information includes the possibility of a high plane position.

[0145] Therefore, for encoding, when entropy encoding is performed on the current plane position which is a low position, the probability of the presence or absence of a plane perpendicular to the Y axis in the current child node is regarded as context information. Among them, this information can be used as plane context information in the plane encoding method and the angle encoding method, and can be included as plane position context information in the plane encoding method.

[0146] In order to Figure 7 (c) Evaluate along the Z axis to determine the plane context information of the current child node 120, which is set as follows:

[0147] If at least one of the child nodes 1, 3, 5, 7 of the first group is occupied and the child nodes 1, 3, 5, 7 of the second group are not occupied, the plane context information comprises a likelihood of a plane and the plane position context information comprises a likelihood of a low plane position.

[0148] If the child nodes 1, 3, 5, 7 of the first group are not occupied and at least one of the child nodes 1, 3, 5, 7 of the second group is occupied, the plane context information comprises a likelihood of a plane and the plane position context information comprises a likelihood of a high plane position.

[0149] ·If the child nodes 1, 3, 5, 7 of the first group are not occupied and the child nodes 1, 3, 5, 7 of the second group are not occupied, the plane context information includes the possibility of the plane, wherein if at least one or more of the child nodes 0, 2, 4, 6 of the adjacent parent node are occupied, the plane position context information includes the possibility of a low plane, otherwise it is "UNKNOWN".

[0150] If at least one of the child nodes 1, 3, 5, 7 of the first group is occupied and at least one of the child nodes 1, 3, 5, 7 of the second group is occupied, and #FirstGroup>#SecondGroup+1, the plane context information includes the possibility of a plane, and the plane position context information includes the possibility of a low plane position.

[0151] If at least one of the child nodes 1, 3, 5, 7 of the first group is occupied and at least one of the child nodes 1, 3, 5, 7 of the second group is occupied, and #FirstGroup<#SecondGroup+1, the plane context information includes the possibility of a plane, and the plane position context information includes the possibility of a high plane position.

[0152] Therefore, for encoding, when entropy encoding is performed on the current plane position which is a low position, the probability of the presence or absence of a plane perpendicular to the Z axis in the current child node is regarded as context information. Among them, this information can be used as plane context information in the plane encoding method and the angle encoding method, and can be included as plane position context information in the plane encoding method.

[0153] As shown above, if the current plane position of the current child node to be encoded is a high plane position, the situation will be different. Then, the encoding order of the adjacent parent node adjacent to the current child node is after the current parent node. In this case, the position of the child node of the adjacent parent node is unknown, and only the position of the adjacent parent node itself can be used to determine the context information.

[0154] In particular, if the current plane position is high, the plane context information and the plane position context information are determined based on the following information: the position of the adjacent parent node immediately behind the current parent node along the axis under consideration, and the position of the group of four child nodes within the current parent node in a plane perpendicular to the axis under consideration and adjacent to the current child node. In other words, the position of the child nodes of the adjacent parent node adjacent to the common surface of the current parent node is usually unknown due to the decoding order. Therefore, the position of the adjacent parent node itself, which shares a common surface with the current child node, is used together with a group arranged on the opposite side of the current parent node (i.e., opposite to the adjacent parent node) within the current parent node. Wherein, the group is preferably defined by four child nodes to form a plane, wherein the plane is perpendicular to the geometric axis under consideration and adjacent to the current child node to be evaluated within the current parent node.

[0155] In general, for the case where the current plane position is low, if at least one child node of the group is occupied and the adjacent parent node is not occupied, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of a low plane position. If the group is not occupied and the adjacent parent node is not occupied, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of a high plane position. If the group is not occupied and the adjacent parent node is not occupied, the plane context information includes the possibility of the plane, wherein if at least one or more child nodes in the adjacent parent node immediately preceding the current parent node along the axis under consideration are occupied, the plane position context information includes the possibility of a low plane, otherwise it is "UNKNOWN". If at least one child node of the group is occupied and the adjacent parent node is occupied, and #Group>2, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of a low plane position. If at least one child node of the group is occupied and the adjacent parent node is occupied, and #Group<2, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of a high plane position. If at least one child node of the group is occupied and the adjacent parent node is occupied, and #Group = 2, the plane context information includes the possibility that there is no plane and therefore the plane position is unknown. Where #Group indicates the number of placeholder child nodes in the group.

[0156] More specifically, for reference Figure 8 (a) Evaluation along the X-axis:

[0157] If at least one of the child nodes 0 to 3 of the group is occupied and the adjacent parent node is not occupied, the plane context information includes a likelihood of a plane and the plane position context information includes a likelihood of a low plane position.

[0158] If child nodes 0 to 3 of the group are unoccupied and the adjacent parent node is occupied, the plane context information includes a likelihood of a plane, and the plane position context information includes a likelihood of a high plane position.

[0159] If child nodes 0 to 3 of the group are not occupied and the adjacent parent node is not occupied, the plane context information includes the possibility of a plane if at least one or more child nodes 4 to 7 (such as Fig. 9 (depicted in (a)) is occupied, the plane location context information includes the possibility of a low plane, otherwise it is "UNKNOWN".

[0160] If at least one of the group's child nodes 0 to 3 is occupied and the adjacent parent node is occupied, and #Group>2, the plane context information includes the likelihood of a plane, and the plane position context information includes the likelihood of a low plane position.

[0161] If at least one of the group's child nodes 0 to 3 is occupied and the adjacent parent node is occupied, and #Group<2, the plane context information includes the likelihood of a plane, and the plane position context information includes the likelihood of a high plane position.

[0162] If at least one of the child nodes 0 to 3 of the group is occupied and the adjacent parent node is occupied, and #Group=2, the plane context information includes the possibility that there is no plane and the plane position is inferred to be unknown. Where #Group indicates the number of placeholder nodes in the group.

[0163] Therefore, for encoding, when entropy encoding is performed on the current plane position which is a low position, the probability of the presence or absence of a plane perpendicular to the X-axis in the current child node is regarded as context information. Among them, this information can be used as plane context information in the plane encoding method and the angle encoding method, and can be included as plane position context information in the plane encoding method.

[0164] In order to Figure 8 (b) Evaluate along the Y axis to determine the plane context information of the current child node 120, which is set as follows:

[0165] If at least one of the child nodes 0, 1, 4, 5 of the group is occupied and the adjacent parent node is not occupied, the plane context information includes a likelihood of a plane and the plane position context information includes a likelihood of a low plane position.

[0166] If child nodes 0, 1, 4, 5 of the group are not occupied and the adjacent parent node is occupied, then the plane context information includes a likelihood of a plane and the plane position context information includes a likelihood of a high plane position.

[0167] If the group's child nodes 0, 1, 4, 5 are not occupied and the adjacent parent node is not occupied, the plane context information includes the possibility of a plane if at least one or more of the adjacent parent nodes 2, 3, 6, 7 (e.g., Fig. 9 (depicted in (b)) is occupied, the plane location context information includes the possibility of a low plane, otherwise it is "UNKNOWN".

[0168] If at least one of the group's child nodes 0, 1, 4, 5 is occupied and the adjacent parent node is occupied, and #Group>2, then the plane context information includes a likelihood of a plane, and the plane position context information includes a likelihood of a low plane position.

[0169] If at least one of the group's child nodes 0, 1, 4, 5 is occupied and the adjacent parent node is occupied, and #Group<2, then the plane context information includes the likelihood of a plane, and the plane position context information includes the likelihood of a high plane position.

[0170] If at least one of the child nodes 0, 1, 4, 5 of the group is occupied and the adjacent parent node is occupied, and #Group = 2, the plane context information includes the possibility that the plane does not exist and the plane position is inferred to be unknown. Where #Group indicates the number of placeholder nodes in the group.

[0171] Therefore, for encoding, when entropy encoding is performed on the current plane position which is a low position, the probability of the presence or absence of a plane perpendicular to the Y axis in the current child node is regarded as context information. Among them, this information can be used as plane context information in the plane encoding method and the angle encoding method, and can be included as plane position context information in the plane encoding method.

[0172] In order to Figure 8 (c) Evaluate along the Z axis to determine the plane context information of the current child node 120, which is set as follows:

[0173] If at least one of the child nodes 0, 2, 4, 6 of the group is occupied and the adjacent parent node is not occupied, the plane context information includes a likelihood of a plane and the plane position context information includes a likelihood of a low plane position.

[0174] If child nodes 0, 2, 4, 6 of the group are not occupied and the adjacent parent node is occupied, then the plane context information includes a likelihood of a plane and the plane position context information includes a likelihood of a high plane position.

[0175] If the group's child nodes 0, 2, 4, 6 are not occupied and the adjacent parent node is not occupied, the plane context information includes the possibility of a plane if at least one or more of the adjacent parent nodes 1, 3, 5, 7 (e.g., Fig. 9 (c) is occupied, the plane location context information includes the possibility of a low plane, otherwise it is "UNKNOWN".

[0176] If at least one of the group's child nodes 0, 2, 4, 6 is occupied and the adjacent parent node is occupied, and #Group>2, then the plane context information includes the likelihood of the plane, and the plane position context information includes the likelihood of the lower plane position.

[0177] If at least one of the group's child nodes 0, 2, 4, 6 is occupied and the adjacent parent node is occupied, and #Group<2, then the plane context information includes the likelihood of a plane, and the plane position context information includes the likelihood of a high plane position.

[0178] If at least one of the child nodes 0, 2, 4, 6 of the group is occupied and the adjacent parent node is occupied, and #Group=2, the plane context information includes the possibility that the plane does not exist and the plane position is inferred to be unknown. Where #Group indicates the number of placeholder nodes in the group.

[0179] Therefore, for encoding, when entropy encoding is performed on the current plane position which is a low position, the probability of the presence or absence of a plane perpendicular to the Y axis in the current child node is regarded as context information. Among them, this information can be used as plane context information in the plane encoding method and the angle encoding method, and can be included as plane position context information in the plane encoding method.

[0180] In summary, in all cases, a reliable plane context information model is built that can be used for entropy coding to account for the occurrence of surfaces or planes in individual nodes and use this knowledge to efficiently predict the occupancy of nodes for data compression by the entropy encoder / decoder.

[0181] In some cases, it is impossible or useless to implement the above-mentioned context information set. Therefore, for the plane position context information in the plane coding mode, if the nearest encoded node at the same coordinate at the same depth is not planar or does not exist, the current GPCC specification recommends the use of a specific context model that uses an axisIdx statement (value = 0, 1 or 2) to identify the axis perpendicular to the plane on which planePosition is encoded. However, this method ignores the information provided by the occupancy of adjacent nodes. According to the present invention, it is still possible to infer based on the occupancy of adjacent child nodes in the same parent node as the current child node and in the same plane perpendicular to the axis under consideration. For example, if the X-axis is considered and the current child node is at coordinate 2, the information of child nodes 0 and 1 in the current parent node and in the same vertical plane perpendicular to the X-axis is known and can be used to determine the plane position context information, such as by using the plane position context information of one of these nodes or most of these nodes as the plane position context information of the current child node to be encoded. For example, if the X-axis is considered and the current child node is at coordinate 7, the information of child nodes 4 to 6 in the same vertical plane perpendicular to the X-axis in the current parent node is known and can be used to determine the plane position context information. Therefore, if there is no recently encoded node at the same coordinates at the same depth D, the plane position context information can replace the proposed axisIdx statement to provide improved context information for more efficient entropy coding. Therefore, the context model for plane position context information in the above-proposed plane coding method is replaced by the following:

[0182] a) The planar position of the current child node according to the occupancy of at least one neighboring node.

[0183] In addition, for the angle encoding method, even if the point cloud is indicated to be encoded by the angle encoding method, one or more nodes of the point cloud may not be suitable for the encoding method. In this case, the plane position context information of the angle encoding method is replaced by the plane position context information according to the plane mode, as described in more detail above.

[0184] Reference now Fig.10 (a), which shows a method for encoding a point cloud, comprising the following steps:

[0185] Determine a coding method, wherein the coding method includes a plane coding method and an angle coding method (step 10);

[0186] Get the encoding context information of the current child node (step 11), where

[0187] If the encoding mode is a plane encoding mode, the encoding context information includes plane context information, wherein the plane context information of the current child node is determined according to the position of the current parent node and the position of at least one adjacent parent node immediately adjacent to the current parent node (step 111);

[0188] If the encoding method is an angle encoding method, the encoding context information of the current child node includes the plane information of the current parent node and the placeholders of the vertically adjacent parent nodes adjacent to the current parent node (step 112); and

[0189] The current child node is entropy encoded based on the determined encoding context information to generate encoding data for a bit stream (step 12).

[0190] Reference now Fig.10 (b), which shows a method for decoding a point cloud, comprising the following steps:

[0191] Determine a coding mode, wherein the coding mode includes a plane coding mode and an angle coding mode (step 20);

[0192] Obtain the encoding context information of the current child node (step 21), where

[0193] If the encoding mode is a plane encoding mode, the encoding context information includes plane context information, wherein the plane context information of the current child node is determined according to the occupancy of the current parent node and the occupancy of at least one adjacent parent node immediately adjacent to the current parent node (step 211);

[0194] If the encoding method is an angle encoding method, the encoding context information of the current child node includes the plane information of the current parent node and the placeholders of the vertically adjacent parent nodes adjacent to the current parent node (step 212); and

[0195] The bitstream is entropy decoded based on the context information of the current child node to reconstruct the point cloud (step 22).

[0196] Therefore, according to the present invention, information about neighboring nodes is used and evaluated to obtain a context information model, thereby more reliably predicting the presence of surfaces within the current child node. This information can be used to reduce the amount of data of the compressed point cloud data, wherein the same context information model is also used for the decoder to be able to efficiently and accurately decode the provided bit stream reliably. Therefore, a significant reduction of at least 1% can be achieved relative to previous encoding methods and current GPCC specifications. However, this value depends on the density of the points, wherein, in particular for indoor point clouds, such as LiDAR point clouds, it is assumed that they can be regarded as dense point clouds.

[0197] In an embodiment of the encoding or decoding method, if the encoding mode is a plane encoding mode, the encoding context information includes plane context information, and the plane context information includes:

[0198] The existence of the plane is determined according to the occupation mode of the current parent node and the occupation mode of at least one adjacent parent node immediately adjacent to the current parent node.

[0199] In an embodiment of the encoding or decoding method, if the encoding mode is a plane encoding mode, the encoding context information includes plane position context information, wherein the plane position context information includes one or more of the following items:

[0200] The distance d to the nearest encoded node at the same coordinates in the corresponding parent node at the same depth D in the octree-based structure;

[0201] The planar position of the most recently encoded node at the same coordinates in the corresponding parent node at the same depth D in the octree-based structure;

[0202] The plane position of the current child node is determined according to the placeholder mode of at least one adjacent parent node immediately adjacent to the current parent node.

[0203] In an embodiment of the encoding or decoding method, if there is no most recently encoded node at the same coordinates in the corresponding parent node at the same depth D in the octree-based structure, the plane position context information only includes the plane position of the current child node determined according to the occupancy pattern of at least one neighboring parent node adjacent to the current parent node.

[0204] In an embodiment of the encoding or decoding method, if the encoding mode is an angle encoding mode, the encoding context information includes plane context information, wherein the plane context information includes one or more of the following items:

[0205] The isPlanar flag of the current parent node; and

[0206] The placeholder for the vertically adjacent parent node to the current parent node (ie, perpendicular to the axis being considered and immediately preceding the current parent node in decoding order).

[0207] In an embodiment of the encoding or decoding method, if the encoding mode is an angle encoding mode and the current child node is not applicable to the angle encoding mode, the encoding context information includes plane position context information, wherein the plane position context information includes one or more of the following items:

[0208] The distance d to the nearest encoded node at the same coordinates in the corresponding parent node at the same depth D in the octree-based structure;

[0209] The planar position of the most recently encoded node at the same coordinates in the corresponding parent node at the same depth D in the octree-based structure;

[0210] The plane position of the current child node is determined according to the placeholder mode of at least one adjacent parent node immediately adjacent to the current parent node.

[0211] In an embodiment of the encoding or decoding method, in order to determine plane context information and plane position context information based on the occupancy pattern of at least one adjacent parent node, the current plane position of the current child node relative to the current parent node is determined, wherein the current plane position includes a high plane position and a low plane position.

[0212] In an embodiment of the encoding or decoding method, if the current plane position is a low plane position, the plane context information and the plane position context information are determined based on: the occupancy of a first group of four child nodes of an adjacent parent node that are adjacent to the current parent node and the occupancy of a second group of four child nodes of the current parent node that are directly opposite to the first group.

[0213] In an embodiment of the encoding or decoding method, for the current plane position being in a low plane position, if at least one child node of the first group is occupied and the second group is not occupied, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of the low plane position.

[0214] In an embodiment of the encoding or decoding method, for the current plane position being in a low plane position, if the first group is not occupied and at least one child node of the second group is occupied, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of a high plane position.

[0215] In an embodiment of the encoding or decoding method, for the current plane position being in a low plane position, if the first group is not occupied and the second group is not occupied, the plane context information includes the possibility of the plane, wherein if at least one or more child nodes in the adjacent parent node that do not belong to the first group are occupied, the plane position context information includes the possibility of the low plane, otherwise it is "UNKNOWN".

[0216] In an embodiment of the encoding or decoding method, for the current plane position being in a low plane position, if at least one child node of the first group is occupied and at least one child node of the second group is occupied, and #FirstGroup>#SecondGroup+1, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of the low plane position, wherein #FirstGroup indicates the number of placeholder nodes in the first group, and #SecondGroup indicates the number of placeholder nodes in the second group.

[0217] In an embodiment of the encoding or decoding method, for the current plane position being in a low plane position, if at least one child node of the first group is occupied and at least one child node of the second group is occupied, and #FirstGroup<#SecondGroup+1, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of a high plane position; wherein #FirstGroup indicates the number of placeholder nodes in the first group, and #SecondGroup indicates the number of placeholder nodes in the second group.

[0218] In an embodiment of the encoding or decoding method, if the current plane position is high, the plane context information and the plane position context information are determined based on: the occupancy of the adjacent parent node immediately behind the current parent node along the axis considered, and the occupancy of the group of four child nodes of the current parent node in a plane perpendicular to the axis considered and immediately adjacent to the current child node.

[0219] In an embodiment of the encoding or decoding method, for the current plane position being in a high plane position, if at least one child node of the group is occupied and the adjacent parent node is not occupied, the plane context information includes a possibility of the plane, and the plane position context information includes a possibility of a low plane position

[0220] In an embodiment of the encoding or decoding method, for the current plane position being in a high plane position, if the group is not occupied and the adjacent parent node is occupied, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of the high plane position

[0221] In an embodiment of the encoding or decoding method, for the current plane position being in a high plane position, if the group is not occupied and the adjacent parent node is not occupied, the plane context information includes the possibility of the plane, wherein if at least one or more child nodes in the adjacent parent node immediately preceding the current parent node along the axis considered are occupied, the plane position context information includes the possibility of a low plane, otherwise it is "UNKNOWN".

[0222] In an embodiment of the encoding or decoding method, for the current plane position being in a high plane position, if at least one child node of the group is occupied and the adjacent parent node is occupied, and #Group>2, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of a low plane position, wherein #Group indicates the number of placeholder nodes in the group.

[0223] In an embodiment of the encoding or decoding method, for the current plane position being in a high plane position, if at least one child node of the group is occupied and the adjacent parent node is occupied, and #Group<2, the plane context information includes the possibility of the plane, and the plane position context information includes the possibility of a high plane position, wherein #Group indicates the number of placeholder nodes in the group.

[0224] In an embodiment of the encoding or decoding method, for the current plane position being at a high plane position, if at least one child node of the group is occupied and the adjacent parent node is occupied, and #Group=2, the plane context information includes the possibility that there is no plane and the plane position is inferred to be unknown; wherein #Group indicates the number of placeholder nodes in the group.

[0225] In an embodiment of the encoding or decoding method, adjacent parent nodes are arranged along one of the geometric axes.

[0226] In an embodiment of the encoding or decoding method, plane context information is determined along all three geometric axes.

[0227] In an embodiment of the encoding or decoding method, the current parent node is arranged relative to the adjacent parent nodes in the order of decoding.

[0228] In a preferred embodiment, a method for encoding / decoding a point cloud to generate a bitstream of compressed point cloud data is implemented in a LIDAR (light detection and ranging) device. The LIDAR device includes a light transmission module and a sensor module. Among them, the light transmission module is configured to scan the environment with a laser, and use the sensor of the sensor module to measure the echo of the laser reflected by the object in the environment. In addition, the LIDAR device includes an evaluation module, which is configured to preferably determine the 3D representation of the environment in the point cloud by the difference in the laser return time and / or the wavelength of the reflected laser. Therefore, the echo can include up to millions of position information points of the object or environment, resulting in a large point cloud, which increases the need for the computing device to further process or evaluate the point cloud. In specific applications such as autonomous driving, the processing of the LIDAR point cloud must be almost real-time due to safety requirements. Therefore, it is necessary to effectively and accurately compress the point cloud data. Therefore, the LIDAR device may include an encoder, which includes a processor and a memory storage device. The memory storage device can store a computer program or application containing instructions, which when executed causes the processor to perform operations such as those described herein. For example, the instructions may encode and output a bitstream encoded according to the method described herein. Additionally or alternatively, the LIDAR device may include a decoder comprising a processor and a memory storage device. The memory storage device may include a computer program or application containing instructions that, when executed, cause the processor to perform operations such as those described herein. Therefore, through the encoder / decoder, point cloud data can be effectively and accurately compressed, thereby providing a probability of more effectively and preferably real-time processing of the acquired point cloud data. Preferably, the processor of the encoder is the same as the processor of the decoder. Preferably, the memory storage device of the encoder is the same as the memory storage device of the decoder. Preferably, the processor of the encoder and / or decoder is also configured to further process or evaluate the point cloud, even more preferably in real time. In particular, for the example of autonomous driving, the evaluation of the point cloud may include determining obstacles in the driving direction.

[0229] Reference now Fig.11, which shows a simplified block diagram of an exemplary embodiment of an encoder 1100. The encoder 1100 includes a processor 1102 and a memory storage device 1104. The memory storage device 1104 can store a computer program or application containing instructions, which when executed causes the processor 1102 to perform operations such as those described herein. For example, the instructions can encode and output a bit stream encoded according to the method described herein. It should be understood that the instructions can be stored on a non-temporary computer-readable medium, such as a compressed disk, a flash memory device, a random access memory, a hard drive, etc. When the instructions are executed, the processor 1102 performs the operations and functions specified in the instructions so as to be used as a dedicated processor to implement the above-mentioned processing. In some examples, such a processor may be referred to as a "processor circuit" or a "processor loop".

[0230] Now also refer to Fig.12 , which shows a simplified block diagram of an exemplary embodiment of a decoder 1200. The decoder 1200 includes a processor 1202 and a memory storage device 1204. The memory storage device 1204 may include a computer program or application containing instructions that, when executed, cause the processor 1202 to perform operations such as those described herein. It should be understood that the instructions may be stored on a computer-readable medium, such as a compressed disk, a flash memory device, a random access memory, a hard drive, etc. When the instructions are executed, the processor 1202 performs the operations and functions specified in the instructions so as to function as a dedicated processor to implement the above-described processes and methods. In some examples, such a processor may be referred to as a "processor circuit" or a "processor loop."

[0231] It should be understood that the decoder and / or encoder according to the present application can be implemented in multiple computing devices, including but not limited to servers, appropriately programmed general-purpose computers, machine vision systems, and mobile devices. The decoder or encoder can be implemented by software containing instructions for configuring one or more processors to perform the functions described herein. The software instructions can be stored on any suitable non-transitory computer-readable memory, including CD, RAM, ROM, flash memory, etc.

[0232] It should be understood that the decoder and / or encoder described herein and the modules, routines, processes, threads or other software components that implement the described methods / processing for configuring encoders or decoders can be implemented using standard computer programming techniques and languages. The application is not limited to specific processors, computer languages, computer programming protocols, data structures, other such implementation details. Those skilled in the art will recognize that the described processing can be implemented as a part of a computer executable code stored in a volatile or non-volatile memory, as a part of an application specific integrated chip (ASIC), etc.

[0233] The present application also provides a computer readable signal encoding data produced by application of an encoding process according to the present application.

[0234] In one aspect of the present invention, a method for encoding a point cloud to generate a bitstream of compressed point cloud data is provided, wherein the geometric structure of the point cloud is represented by an octree-based structure, the octree-based structure comprising a plurality of nodes having a parent-child relationship by recursively partitioning a volume space containing the point cloud into sub-volumes, each sub-volume being associated with a node of the octree-based structure, the method comprising the following steps:

[0235] Determine the encoding method;

[0236] Get the encoding context information of the current child node;

[0237] The placeholder of the current child node is entropy encoded based on the determined encoding context information to generate encoding data for the bitstream.

[0238] Preferably, the encoding method includes a plane encoding method and / or an angle encoding method.

[0239] Preferably, if the encoding mode is a plane encoding mode, the encoding context information includes plane context information.

[0240] Preferably, the plane context information of the current child node is determined according to the placeholder mode of the current parent node and the placeholder mode of at least one adjacent parent node immediately adjacent to the current parent node;

[0241] Preferably, if the encoding method is an angle encoding method, the encoding context information of the current child node includes the plane information of the current parent node and the placeholders of the vertically adjacent parent nodes adjacent to the current parent node.

[0242] In one aspect of the present invention, a method for decoding a bit stream of compressed point cloud data to generate a reconstructed point cloud, wherein the geometric structure of the point cloud is represented by an octree-based structure, the octree-based structure comprising a plurality of nodes having a parent-child relationship by recursively partitioning a volume space containing the point cloud into sub-volumes, each sub-volume being associated with a node of the octree-based structure, the method comprising the following steps:

[0243] Determine the encoding method;

[0244] Get the encoding context information of the current child node;

[0245] The bitstream is entropy decoded based on the encoded context information of the current child node to reconstruct the point cloud.

[0246] Preferably, the encoding method includes a plane encoding method and / or an angle encoding method.

[0247] Preferably, if the encoding mode is a plane encoding mode, the encoding context information includes plane context information.

[0248] Preferably, the plane context information of the current child node is determined according to the placeholder pattern of the current parent node and the placeholder pattern of at least one adjacent parent node immediately adjacent to the current parent node.

[0249] Preferably, if the encoding method is an angle encoding method, the encoding context information of the current child node includes the plane information of the current parent node and the placeholders of the vertically adjacent parent nodes adjacent to the current parent node.

[0250] Certain adjustments and modifications may be made to the described embodiments. Therefore, the embodiments discussed above are considered to be illustrative rather than restrictive. In particular, the embodiments may be freely combined with each other.

Claims

1. A method for encoding a point cloud to generate a bitstream of compressed point cloud data, wherein: The geometric structure of the point cloud is represented by an octree-based structure, the octree-based structure comprising a plurality of nodes having a parent-child relationship by recursively partitioning a volume space containing the point cloud into sub-volumes, each of the sub-volumes being associated with a node of the octree-based structure, the method comprising the following steps: Determine a coding mode, wherein the coding mode includes a plane coding mode and an angle coding mode; Get the encoding context information of the current child node, where If the encoding mode is the plane encoding mode, the encoding context information includes plane context information, wherein the plane context information of the current child node is determined according to the placeholder mode of the current parent node and the placeholder mode of at least one adjacent parent node adjacent to the current parent node; wherein the placeholder mode includes plane information of the probability of whether the node is occupied; The coding context information further includes plane position context information, wherein the plane position context information includes one or more of the following items: the distance d to the nearest encoded node at the same coordinates in the corresponding parent node at the same depth D in the octree-based structure; The planar position of the nearest encoded node at the same coordinates in the corresponding parent node at the same depth D in the octree-based structure; The plane position of the current child node is determined according to the placeholder pattern of at least one adjacent parent node immediately adjacent to the current parent node; If the encoding method is the angle encoding method, the encoding context information of the current child node includes the plane information of the current parent node and the placeholder of the vertically adjacent parent node adjacent to the current parent node, Performing entropy coding on the current child node based on the determined coding context information to generate coded data for the bitstream; In order to determine the plane context information and the plane position context information according to the occupancy mode of the at least one adjacent parent node, a current plane position of the current child node relative to the current parent node is determined, wherein the current plane position includes a high plane position and a low plane position; Wherein, if the current plane position is a low plane position, the plane context information and the plane position context information are determined according to: The placeholders of the first group of four child nodes of the adjacent parent node immediately adjacent to the current parent node, and The placeholders of the second group of four child nodes of the current parent node that are opposite to the first group; Wherein, if the current plane position is high, the plane context information and the plane position context information are determined according to: the placeholder of the adjacent parent node immediately following the current parent node along the considered axis, and The placeholder of the group of four child nodes of the current parent node in the plane perpendicular to the considered axis and immediately adjacent to the current child node.

2. A method for decoding a bit stream of compressed point cloud data to generate a reconstructed point cloud, wherein: The geometric structure of the point cloud is represented by an octree-based structure, the octree-based structure comprising a plurality of nodes having a parent-child relationship by recursively partitioning a volume space containing the point cloud into sub-volumes, each of the sub-volumes being associated with a node of the octree-based structure, the method comprising the following steps: Determine a coding mode, wherein the coding mode includes a plane coding mode and an angle coding mode; Get the encoding context information of the current child node, where If the encoding mode is the plane encoding mode, the encoding context information includes plane context information, wherein the plane context information of the current child node is determined according to the placeholder mode of the current parent node and the placeholder mode of at least one adjacent parent node adjacent to the current parent node; wherein the placeholder mode includes plane information of the probability of whether the node is occupied; The coding context information further includes plane position context information, wherein the plane position context information includes one or more of the following items: the distance d to the nearest encoded node at the same coordinates in the corresponding parent node at the same depth D in the octree-based structure; The planar position of the nearest encoded node at the same coordinates in the corresponding parent node at the same depth D in the octree-based structure; The plane position of the current child node is determined according to the placeholder pattern of at least one adjacent parent node immediately adjacent to the current parent node; If the encoding mode is the angle encoding mode, the encoding context information of the current child node includes the plane information of the current parent node and the placeholder of the vertically adjacent parent node adjacent to the current parent node; and Performing entropy decoding on the bitstream based on the encoding context information of the current child node to reconstruct the point cloud; In order to determine the plane context information and the plane position context information according to the occupancy mode of the at least one adjacent parent node, a current plane position of the current child node relative to the current parent node is determined, wherein the current plane position includes a high plane position and a low plane position; Wherein, if the current plane position is a low plane position, the plane context information and the plane position context information are determined according to: The placeholders of the first group of four child nodes of the adjacent parent node immediately adjacent to the current parent node, and The placeholders of the second group of four child nodes of the current parent node that are opposite to the first group; Wherein, if the current plane position is high, the plane context information and the plane position context information are determined according to: the placeholder of the adjacent parent node immediately following the current parent node along the considered axis, and The placeholder of the group of four child nodes of the current parent node in the plane perpendicular to the considered axis and immediately adjacent to the current child node.

3. The method according to claim 1 or 2, characterized in that If there is no most recently encoded node at the same coordinates in the corresponding parent node at the same depth D in the octree-based structure, the plane position context information only includes the plane position of the current child node determined according to the occupancy pattern of at least one neighboring parent node adjacent to the current node.

4. The method according to claim 1 or 2, characterized in that: If the encoding mode is the angle encoding mode and the current child node is not applicable to the angle encoding mode, the encoding context information further includes plane position context information, wherein the plane position context information includes one or more of the following items: the distance d to the nearest encoded node at the same coordinates in the corresponding parent node at the same depth D in the octree-based structure; The planar position of the nearest encoded node at the same coordinates in the corresponding parent node at the same depth D in the octree-based structure; The plane position of the current child node is determined according to the occupation mode of at least one adjacent parent node immediately adjacent to the current parent node.

5. The method according to claim 1 or 2, characterized in that: If at least one child node of the first group is occupied and the second group is not occupied, the plane context information includes a likelihood probability parameter of a plane, and the plane position context information includes a likelihood probability parameter of a low plane position; or If the first group is not occupied and at least one child node of the second group is occupied, the plane context information includes a likelihood probability parameter for a plane, and the plane position context information includes a likelihood probability parameter for a high plane position; or If the first group is not occupied and the second group is not occupied, the plane context information includes a possibility probability parameter of a plane, wherein if at least one or more child nodes of the adjacent parent node that do not belong to the first group are occupied, the plane position context information includes a possibility probability parameter of a low plane; or If at least one child node of the first group is occupied and at least one child node of the second group is occupied, and #FirstGroup>#SecondGroup+1, the plane context information includes a possibility probability parameter of a plane, and the plane position context information includes a possibility probability parameter of a low plane position; or If at least one child node of the first group is occupied and at least one child node of the second group is occupied, and #FirstGroup<#SecondGroup+1, the plane context information includes a likelihood probability parameter of a plane, and the plane position context information includes a likelihood probability parameter of a high plane position; Among them, #FirstGroup indicates the number of placeholder child nodes in the first group, and #SecondGroup indicates the number of placeholder child nodes in the second group.

6. The method according to claim 1 or 2, characterized in that: If at least one child node of the group is occupied and the adjacent parent node is not occupied, the plane context information includes a likelihood probability parameter of the plane, and the plane position context information includes a likelihood probability parameter of a low plane position; or If the group is not occupied and the adjacent parent node is occupied, the plane context information includes a likelihood probability parameter of a plane, and the plane position context information includes a likelihood probability parameter of a high plane position; or If the group is not occupied and the adjacent parent node is not occupied, the plane context information includes a likelihood probability parameter for a plane, wherein the plane position context information includes a likelihood probability parameter for a low plane if at least one or more child nodes in the adjacent parent node immediately preceding the current parent node along the considered axis are occupied; or If at least one child node of the group is occupied and the adjacent parent node is occupied, and #Group>2, the plane context information includes a likelihood probability parameter of a plane, and the plane position context information includes a likelihood probability parameter of a low plane position; or If at least one child node of the group is occupied and the adjacent parent node is occupied, and #Group<2, the plane context information includes a likelihood probability parameter of a plane, and the plane position context information includes a likelihood probability parameter of a high plane position; or If at least one child node of the group is occupied and the adjacent parent node is occupied, and #Group=2, the plane context information includes a probability parameter of the possibility that a plane does not exist and the plane position is inferred to be unknown; Here, #Group indicates the number of placeholder nodes in the group.

7. The method according to claim 1 or 2, characterized in that: The neighboring parent nodes are arranged along one of the considered geometric axes.

8. The method according to claim 1 or 2, characterized in that: The context information is determined along all three geometric axes.

9. An encoder for encoding a point cloud to generate a bit stream of compressed point cloud data, wherein: The geometric structure of the point cloud is represented by an octree-based structure, the octree-based structure comprising a plurality of nodes having a parent-child relationship by recursively partitioning a volume space containing the point cloud into sub-volumes, each of the sub-volumes being associated with a node of the octree-based structure, the encoder comprising: processor, and A memory storage device, wherein the memory storage device stores instructions executable by the processor, and when the instructions are executed, the processor is prompted to perform the method of claim 1 and any one of claims 3 to 8 dependent on claim 1.

10. A decoder for decoding a bit stream of compressed point cloud data to generate a reconstructed point cloud, wherein: The geometric structure of the point cloud is represented by an octree-based structure, the octree-based structure comprising a plurality of nodes having a parent-child relationship by recursively partitioning a volume space containing the point cloud into sub-volumes, each of the sub-volumes being associated with a node of the octree-based structure, the decoder comprising: processor, and A memory storage device, wherein the memory storage device stores instructions executable by the processor, and when the instructions are executed, the processor is prompted to execute the method of claim 2 and any one of claims 3 to 8 dependent on claim 2.

11. A non-transitory computer-readable storage medium storing instructions for execution by a processor, which, when executed by the processor, causes the processor to execute the method according to any one of claims 1 to 8.