WO2022120542A1 - Procédé et appareil de codage de nuage de points, procédé et appareil de décodage de nuage de points, et support de stockage lisible par ordinateur - Google Patents

Procédé et appareil de codage de nuage de points, procédé et appareil de décodage de nuage de points, et support de stockage lisible par ordinateur Download PDF

Info

Publication number
WO2022120542A1
WO2022120542A1 PCT/CN2020/134355 CN2020134355W WO2022120542A1 WO 2022120542 A1 WO2022120542 A1 WO 2022120542A1 CN 2020134355 W CN2020134355 W CN 2020134355W WO 2022120542 A1 WO2022120542 A1 WO 2022120542A1
Authority
WO
WIPO (PCT)
Prior art keywords
node
occupancy information
sub
nodes
point cloud
Prior art date
Application number
PCT/CN2020/134355
Other languages
English (en)
Chinese (zh)
Inventor
陈嘉枫
虞露
李璞
郑萧桢
Original Assignee
浙江大学
深圳市大疆创新科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 浙江大学, 深圳市大疆创新科技有限公司 filed Critical 浙江大学
Priority to PCT/CN2020/134355 priority Critical patent/WO2022120542A1/fr
Priority to CN202080081330.1A priority patent/CN114885617A/zh
Publication of WO2022120542A1 publication Critical patent/WO2022120542A1/fr

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/90—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using coding techniques not provided for in groups H04N19/10-H04N19/85, e.g. fractals
    • H04N19/96—Tree coding, e.g. quad-tree coding

Definitions

  • the present application relates to the technical field of point cloud data processing, and in particular, to a point cloud encoding method, device, and computer-readable storage medium, and a point cloud decoding method, device, and computer-readable storage medium.
  • a point cloud is a collection of multiple point cloud points in three-dimensional space.
  • the point cloud data of each point cloud point includes geometric data and attribute data. Since a frame of point cloud often contains hundreds of thousands or even hundreds of millions of point cloud points, the scale of point cloud data is very large, and it is necessary to encode and compress point cloud data.
  • the geometric data and attribute data of the point cloud are encoded and decoded separately. Since the encoding and decoding of the attribute data requires the use of the geometric data, the encoding and decoding of the geometric data is usually completed before the encoding and decoding of the attribute data. However, the current encoding and decoding methods of geometric data do not match the encoding and decoding methods of attribute data, which leads to reordering of point clouds after the encoding and decoding of geometric data is completed, which consumes a lot of time and results in poor implementation of the overall solution.
  • the embodiments of the present application provide a point cloud encoding method and a point cloud decoding method, one of the purposes is to solve the need to perform geometric encoding and decoding on the reconstructed point cloud points before performing attribute encoding and decoding on the point cloud.
  • Technical issues with reordering are to solve the need to perform geometric encoding and decoding on the reconstructed point cloud points before performing attribute encoding and decoding on the point cloud.
  • a first aspect of the embodiments of the present application provides a point cloud encoding method, including:
  • the first node is divided into a plurality of sub-nodes, and the first node is any non-leaf node containing point cloud points obtained by tree-like division of the point cloud;
  • the space occupation code is encoded.
  • a second aspect of the embodiments of the present application provides a point cloud decoding method, including:
  • the first node is divided into a plurality of sub-nodes, and the first node is any non-leaf node that includes a point cloud point obtained by tree-like division of the point cloud;
  • the target child node is a non-leaf node, perform the tree division on the target child node.
  • a third aspect of an embodiment of the present application provides a point cloud encoding device, including: a processor and a memory storing a computer program, where the processor implements the following steps when executing the computer program:
  • the first node is divided into a plurality of sub-nodes, and the first node is any non-leaf node containing point cloud points obtained by tree-like division of the point cloud;
  • the space occupation code is encoded.
  • a fourth aspect of the embodiments of the present application provides a point cloud decoding device, comprising: a processor and a memory storing a computer program, where the processor implements the following steps when executing the computer program:
  • the first node is divided into a plurality of sub-nodes, and the first node is any non-leaf node containing point cloud points obtained by tree-like division of the point cloud;
  • the target child node is a non-leaf node, perform the tree division on the target child node.
  • a fifth aspect of the embodiments of the present application provides a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, any point cloud encoding provided by the embodiments of the present application is implemented method.
  • a sixth aspect of an embodiment of the present application provides a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, any point cloud decoding provided by the embodiments of the present application is implemented. method.
  • the space occupation code corresponding to the first node generated by the encoding end is obtained by arranging the occupation information of each sub-node corresponding to the first node in the order of Hilbert order, so that , the decoding end also needs to arrange the sub-nodes in the order of Hilbert order, so that the target sub-nodes containing point cloud points in each sub-node can be determined according to the space occupancy code.
  • the point cloud points have been arranged in the order of the Hilbert order, and there is no need to reorder the point cloud points, which greatly facilitates the follow-up based on the Hilbert order.
  • the encoding and decoding of point cloud attributes improves the overall achievability of the point cloud encoding and decoding scheme.
  • FIG. 1 is a two-dimensional and three-dimensional model diagram of a Hilbert curve provided by an embodiment of the present application.
  • FIG. 2 is a flowchart of a point cloud encoding method provided by an embodiment of the present application.
  • FIG. 3 is a schematic diagram of a division manner of an octree provided by an embodiment of the present application.
  • FIG. 4 is a schematic diagram of tree-like division of a point cloud provided by an embodiment of the present application.
  • FIG. 5A to FIG. 5C are the Hilbert order-based distances from the first child node of 1, The distribution map of the surrounding child nodes of .
  • FIG. 6 is a schematic diagram of an encoding and decoding sequence of node occupancy information based on Morton sequence provided by an embodiment of the present application.
  • FIG. 7 is a distribution diagram of surrounding sub-nodes of a first sub-node based on a Morton order provided by an embodiment of the present application.
  • FIG. 8 is a flowchart of a point cloud decoding method provided by an embodiment of the present application.
  • FIG. 9 is a schematic structural diagram of a point cloud encoding apparatus provided by an embodiment of the present application.
  • FIG. 10 is a schematic structural diagram of a point cloud decoding apparatus provided by an embodiment of the present application.
  • a point cloud is a collection of multiple point cloud points in three-dimensional space.
  • the point cloud data of each point cloud point can include a geometric part and an attribute part.
  • the data of the geometric part can be called geometric data
  • the data of the attribute part can be called attribute data.
  • the geometric data can be used to characterize the position of the point cloud point in the three-dimensional space, for example, it can be the geometric coordinates of the point cloud point
  • the attribute data can include the attributes of the point cloud point in various aspects, for example, it can include the point cloud point Properties in terms of color, reflectivity, etc.
  • a point cloud frame often contains hundreds of thousands or even hundreds of millions of point cloud points, so the scale of point cloud data is very large.
  • point cloud data In order to facilitate the storage and transmission of point cloud data, it is necessary to encode and compress point cloud data.
  • the geometric data and attribute data of the point cloud are encoded and decoded separately. Since the encoding and decoding of the attribute data requires the use of the geometric data, the encoding and decoding of the geometric data is usually completed before the encoding and decoding of the attribute data.
  • the encoding and decoding of point cloud attributes can be performed based on the Hilbert order. Specifically, after determining the Hilbert order corresponding to the plurality of point cloud points, the attribute encoding and decoding of the point cloud points may be sequentially performed according to the order of the point cloud points on the Hilbert order. Since the Hilbert order has good spatial neighbor characteristics, the point cloud points that are adjacent or similar in order to the currently encoded point cloud point (hereinafter referred to as the current point) will also be close to the current point in spatial distance, so , the prediction of the attribute value of the current point based on the attribute value of the point cloud point that has been encoded and decoded before the sequence can be more accurate, and the compression performance can be improved. Referring to FIG. 1 , FIG. 1 is a two-dimensional and three-dimensional model diagram of a Hilbert curve provided by an embodiment of the present application.
  • the encoding and decoding of the point cloud attributes is performed after the encoding and decoding of the point cloud geometry is completed, and since the reconstructed point cloud points are arranged in Morton order after the encoding and decoding of the geometric data of the point cloud is completed, if using the Greek To perform attribute encoding and decoding on the point cloud in the order of the Albert order, it is necessary to reorder the reconstructed point cloud points after geometric encoding and decoding.
  • the number of reconstructed point clouds is huge, and it takes a lot of time to reorder them, resulting in poor achievability of the overall solution.
  • the embodiment of the present application provides a point cloud encoding method.
  • a flowchart of the point cloud encoding method provided by the embodiment of the present application. The method includes the following steps:
  • the occupancy information is used to indicate whether the child node contains the point cloud point
  • the first node is any non-leaf node containing point cloud points obtained by tree-like division of the point cloud, and the occupancy information of the child node is used to indicate whether the child node contains point cloud points.
  • the point cloud When encoding and decoding the geometric data of the point cloud, in one embodiment, it can be implemented by performing tree division on the point cloud.
  • the point cloud is divided into a tree. Specifically, all the point cloud points in the point cloud can be wrapped by a bounding box, and the bounding box can be recursively divided in a specified way.
  • the so-called recursive division may be that after each division of a node, sub-nodes including point cloud points among the sub-nodes obtained by division are further divided.
  • a node may correspond to a space.
  • the above-mentioned bounding box may also correspond to a node, and the node may be called a root node.
  • the division of nodes is to divide the space.
  • each sub-node corresponds to a small space.
  • the final divided sub-nodes will correspond to a minimum unit space, and the nodes corresponding to the unit space can be called leaf nodes . Since the leaf node contains only one point cloud point, the geometric coordinates of the point cloud point contained in the leaf node can be determined according to the geometric coordinates of the leaf node (for example, the coordinates of the geometric center).
  • FIG. 3 is a A schematic diagram of a division manner of an octree provided by an embodiment of the present application.
  • different division methods may also be used for different divisions, or several division methods may be mixed and used according to certain rules.
  • the encoding end since the real geometric coordinates of each point cloud point in the point cloud are known, when the point cloud is divided into a tree, it can be divided into multiple subgroups according to the real geometric coordinates of the point cloud point. In the node, the child nodes containing the point cloud points are determined, so that these child nodes can be further divided until they are divided into the level of leaf nodes.
  • the geometric coordinates of the point cloud points can be reconstructed according to the geometric coordinates of the leaf nodes, without the need for transmission.
  • the geometric coordinates themselves with a huge amount of data achieve the effect of compression.
  • the decoding end does not know the geometric coordinates of each point cloud point like the encoding end. Therefore, it needs some information to ensure synchronization with the tree division of the encoding end. Here, this information can be called division information.
  • the division information can include the geometric information of the bounding box, and the geometric information of the bounding box can make the bounding box uniquely determined.
  • the geometric information of the bounding box may be the length, width, and height of the bounding box and the geometric coordinates of the center or other vertices.
  • the coordinates determine the shape of the bounding box, and the tree-like division depth of the bounding box in each coordinate direction can be determined according to the maximum and minimum geometric coordinates of the point cloud points.
  • the division information may further include a space occupation code, and the space occupation code may be used to indicate which child nodes are child nodes containing point cloud points to be divided.
  • the encoding end can generate a space occupation code corresponding to the first node after dividing the first node into multiple sub-nodes, and can encode the space occupation code and write it into the code stream.
  • the decoding end can start from The space occupancy code is obtained by decoding the code stream, and according to the space occupancy code, it is determined which of the multiple sub-nodes corresponding to the first node are the sub-nodes that contain point cloud points to be divided, so that these sub-nodes containing point cloud points can be analyzed. further division.
  • the first node may be any non-leaf node containing point cloud points
  • the space occupation code corresponding to the first node in one embodiment, may be composed of each child node corresponding to the first node
  • the occupancy information is arranged in a certain order.
  • the occupancy information of a node can be used to indicate whether the node contains a point cloud point. Since the point cloud point is included and the point cloud point is not included are two different situations, so corresponding to different situations, the occupancy information can be represented in different ways.
  • the occupancy information may be represented by binary symbols 0 and 1. For example, in an example, the first binary symbol 1 may be used to indicate that the node contains point cloud points, and the second binary symbol 0 may be used to indicate that the node contains a point cloud. Indicates that the node does not contain point cloud points.
  • the point cloud When the point cloud is divided into a tree, the point cloud can be gradually divided from the level of the root node to the level of the leaf node to obtain a tree structure.
  • the division order of these nodes to be divided can be based on their Hilber in this layer. The ordering on the special order is determined. For example, after dividing a node, you can get a total of 8 child nodes a, b, c, d, e, f, g, and h, of which, the child nodes c, d, and e contain point cloud points.
  • the 8 sub-nodes can be arranged in Hilbert order.
  • the division order of c, d, and e can be determined according to the order of the sub-nodes, that is, the sub-nodes c, d, and e can be divided in the order of d, e, and c.
  • Each divided node has a corresponding space occupation code, and these space occupation codes can follow certain rules when writing the code stream.
  • the writing order of the space occupation code can correspond to the division order of the nodes. Since the division of nodes by the coding end is carried out layer by layer, that is, only the nodes of the current level are divided, the division of the nodes of the next level will be entered. Therefore, in the code stream, the space occupied by the nodes of the same level Codes can be consecutive.
  • the writing order of the space occupancy codes of the nodes at the same level also corresponds to the division order of the nodes in the level.
  • the space occupancy codes corresponding to e will be continuous in the code stream, and since the sub-nodes c, d and e are divided in the order d, e, c, the space occupancy codes corresponding to the sub-nodes c, d and e are also in accordance with d, d, and e.
  • the sequence of e and c is written into the code stream. Referring to the example shown in FIG. 4 , the sequence of the space occupancy codes written in the code stream can be determined as follows: 00000010, 00010010, 00110000, 00011010 according to the overall division order of layer-by-layer downward and single-layer from left to right.
  • the encoding end divides the first node into multiple sub-nodes and determines the occupancy information of each sub-node, and then converts the occupancy information of each sub-node into Morton order. Sequentially arranged to generate space occupation codes.
  • the decoding end can decode the space occupancy code from the code stream, but when determining the child nodes containing the point cloud points according to the space occupancy code, it is necessary to arrange the child nodes in Morton order first, so as to obtain the space occupancy code.
  • the target child node to be divided in each child node can be determined according to the space occupation code.
  • each point cloud point will be arranged in the order of Morton order, which is not conducive to the subsequent encoding and decoding of point cloud attributes based on the Hilbert order. Reorder.
  • the space occupation code corresponding to the first node generated by the encoding end is obtained by arranging the occupation information of each child node corresponding to the first node in the order of Hilbert order, Therefore, the decoding end also needs to arrange each child node in the order of Hilbert order, so that the target child node including the point cloud point in each child node can be determined according to the space occupation code.
  • the point cloud points after completing the encoding and decoding of the point cloud geometric data, the point cloud points have been arranged in the order of the Hilbert order, and there is no need to reorder the point cloud points, which greatly facilitates the follow-up based on the Hilbert order.
  • the encoding and decoding of point cloud attributes improves the overall achievability of the point cloud encoding and decoding scheme.
  • the encoding of the space occupancy code by the encoding end may have various implementations.
  • the encoder can use the context-based adaptive binary arithmetic coding to encode the space occupancy code, and correspondingly, the decoder can also use the context-based adaptive binary arithmetic decoding to decode the space occupancy code from the code stream. code.
  • the context-based adaptive binary arithmetic coding and decoding encodes and decodes the occupancy information in the space occupancy code one by one. Therefore, for the convenience of understanding, the first occupancy information can be used as the coding and decoding object for description.
  • the first occupancy information is the occupancy information of the first child node, and the first child node is any child node of a plurality of child nodes obtained by dividing the first node.
  • a context model corresponding to the first occupancy information may be determined first, and then the first occupancy information may be arithmetically encoded according to the context model. decoding.
  • the context model may also be referred to as a probability model, and the context model may include two probability values, a first probability value and a second probability value, wherein the first probability value may be that the first occupancy information is the occupancy that includes point cloud points
  • the probability of the information for example, the first probability value can be the probability that the first occupancy information is binary symbol 0, and the second probability value can be the probability that the first occupancy information is the occupancy information that does not contain point cloud points, for example, the second probability value can be is the probability that the first occupancy information is a binary symbol 1.
  • the first occupancy information may be arithmetically encoded and decoded according to the probability.
  • the context model corresponding to the first occupancy information may be determined according to the spatial position of the first child node in the first node.
  • the spatial position of the first child node in the first node may have N situations, and here, the specific value of N may be determined according to the division method of the first node. For example, if the first node is divided by an octree, the first node will be divided into 8 child nodes, then the first child node can be any one of the 8 child nodes, and the first child node is in The spatial position in the first node can have 8 cases.
  • 8 context models can be pre-built corresponding to these 8 situations, then when the context model of the first occupancy information is determined, one corresponding to the spatial position of the first child node can be determined from the 8 context models.
  • the context model of is used as the context model corresponding to the first occupancy information.
  • the context model corresponding to the first occupancy information may be determined according to the occupancy information of the surrounding child nodes of the first child node.
  • the surrounding sub-nodes are sub-nodes that are close to the first sub-node in spatial distance, which may specifically be sub-nodes whose distance from the first sub-node is a first preset value, and the first preset value may include one of the following or multiple values: 1,
  • the distance between the sub-nodes described here is based on the side length of the sub-node as the unit distance.
  • the cube at the center position corresponds to the first sub-node, which is the same as the first sub-node.
  • the surrounding sub-nodes whose node distance is 1 may include adjacent sub-nodes located in 6 directions of front, rear, left, right, up and down of the first sub-node.
  • the distance from the first child node is The surrounding child nodes can include 4 child nodes on the XY plane with the first child node, 4 child nodes on the XZ plane with the first child node, and 4 child nodes on the YZ plane with the first child node, a total of 12 distance is surrounding child nodes.
  • the distance from the first child node is The surrounding child nodes of can have 8, which are located at the 8 vertex corners of the combination of 3*3*3 cubes centered on the first child node.
  • multiple surrounding child nodes corresponding to the first child node may be grouped to obtain multiple child node groups, and a context model corresponding to the first occupancy information may be determined according to the occupancy information corresponding to each child node group .
  • a sub-node group may include at least one sub-node. If a sub-node group contains only one surrounding sub-node, the occupancy information corresponding to the sub-node group may be the occupancy information of the surrounding sub-node;
  • the group contains two or more surrounding child nodes, in one example, when each surrounding child node in the group does not contain point cloud points, it can be determined that the occupancy information of the child node group does not contain point cloud points.
  • Occupancy information if any surrounding sub-node in the sub-node group contains point cloud points, it can be determined that the occupation information corresponding to the sub-node group is the occupation information including point cloud points.
  • the occupancy information including point cloud points is represented by the first binary symbol 1
  • the occupancy information not including point cloud points is represented by the second binary symbol 0
  • An OR operation may be performed on the occupancy information of each surrounding child node in the group, and the operation result of the OR operation may be used as the occupancy information corresponding to the child node group.
  • the surrounding sub-nodes are divided into K sub-node groups in total, since the occupancy information of each sub-node group corresponds to two possibilities (that is, it may contain point cloud points or may not contain point cloud points), therefore, all sub-nodes are integrated.
  • Context models may be established separately for the 2 K situations in advance, and 2 K context models are obtained, wherein each context model may correspond to one situation. In this way, after determining the occupancy information corresponding to each sub-node group, one context model may be determined from the 2K context models as the context model corresponding to the first occupancy information according to the occupancy information corresponding to each sub-node group.
  • the first index may be determined according to the occupancy information corresponding to each sub-node group, and the context corresponding to the first index The model is determined as a context model corresponding to the first occupancy information.
  • the value corresponding to the binary symbol string obtained by arranging the occupancy information corresponding to each sub-node group may be used as the first index.
  • the occupancy information can be represented by binary symbols, so the arrangement of the occupancy information can form a binary symbol string.
  • the binary symbol string obtained by arranging the occupancy information corresponding to the 7 sub-node groups can be 1111000, and the value corresponding to 1111000 is 120, that is, 2 7
  • the context model whose index is 120 in the context models is determined as the context model corresponding to the first occupancy information.
  • the surrounding child nodes may include 6 surrounding child nodes whose distance from the first child node is 1, and the distance from the first child node is The 12 surrounding child nodes of , and the distance from the first child node is 8 surrounding child nodes. In one embodiment, six surrounding child nodes with a distance of 1 may be divided into 3 groups.
  • two surrounding child nodes that are in the same X-axis direction as the first child node may be grouped into one group (that is, the front and rear two A group of child nodes is a group), and the two surrounding child nodes that are in the same Y-axis direction as the first child node can be a group (that is, two left and right child node groups are a group), and the first child node is in the same Z-axis direction.
  • the two surrounding sub-nodes of can be a group (that is, the upper and lower sub-node groups are a group); the distance can be set as The 12 surrounding sub-nodes of the is a group, and the four surrounding sub-nodes that are in the same YZ plane as the first sub-node can be a group; the distance can be set as The 8 surrounding sub-nodes of , are divided into 1 group, so that 7 sub-node groups can be obtained.
  • the above manner of dividing the surrounding sub-nodes into 7 sub-node groups can be understood in conjunction with FIG. 5A to FIG. 5C .
  • the distribution characteristics of point clouds in different directions and on different planes can be taken into account when context modeling, so that in the statistical sense, the probability in the context model corresponding to various situations can be more accurate. Close to the true probability, the compression performance can be improved, and the probability distribution of each context model is more stable, the information entropy redundancy is removed, and the performance of entropy coding is improved.
  • the first preset value can be adjusted according to the actual situation, for example, the first preset value can only include 1 and does not include Therefore, there can be only 18 surrounding child nodes (6 with a distance of 1, and a distance of 12), the first preset value can also include only 1, not including and Therefore, there can be only 6 surrounding child nodes.
  • the grouping manner of the surrounding sub-nodes can also be selected according to actual needs, which is not limited in this embodiment of the present application.
  • the surrounding sub-nodes of the first sub-node do not contain point cloud points. Therefore, in an embodiment, in order to further improve the probability in the context model The accuracy of , can be further classified for the case that none of the surrounding sub-nodes contain point cloud points. Specifically, if the occupancy information of each surrounding child node is the occupancy information indicating that the point cloud point is not included, for example, it is the second binary symbol 0, then the occupancy information of the first node (the parent node of the first child node) can be The occupancy information of the surrounding nodes determines a context model corresponding to the first occupancy information.
  • the surrounding nodes of the first node may be surrounding nodes whose distance from the first node is a second preset value.
  • the second preset value may be 1, and the surrounding nodes may include nodes located in front, rear, left, and right of the first node. Adjacent nodes in six directions: right, up, and down.
  • the second preset value can also include and other values. It should be noted that the distance between nodes described here is based on the side length of the node as the unit distance, which is different from the aforementioned first preset value where the side length of the child node is the unit distance.
  • the number of context models to be constructed may be determined according to the number of surrounding nodes.
  • the second preset value may be 1
  • the first node may include 6 surrounding nodes with a distance of 1. Since the occupancy information of each surrounding node has two possible situations, there are 2 possible situations in total. 6 kinds of context models can be pre-established for each of the 26 kinds of situations, then 26 context models can be obtained.
  • the second index may be determined according to the occupancy information of each surrounding node, so that the corresponding second index can be The context model is determined as the context model corresponding to the first occupancy information.
  • the occupancy information of the surrounding nodes may be a binary symbol, so the arrangement of the occupancy information of the surrounding nodes may form a binary symbol string, and the value corresponding to the binary symbol string may be the second index.
  • the binary symbol string obtained by arranging the occupancy information of the 6 surrounding nodes can be 111100, and the value corresponding to 111100 is 60, That is, the context model with an index of 60 may be determined from the 26 context models as the context model corresponding to the first occupancy information.
  • the context model After determining the context model corresponding to the first occupancy information, for the encoding end, the context model can be used to perform arithmetic coding on the first occupancy information and write it into the code stream.
  • the first occupancy information is obtained by decoding in the middle.
  • FIG. 6 is a schematic diagram of an encoding and decoding sequence of node occupancy information based on Morton sequence provided by an embodiment of the present application. It can be seen that in the example shown in FIG. 6 , the codec order based on the Morton sequence is fixed in the order of 000, 001, 010, 011, 100, 101, 110, and 111.
  • the node located in the positive direction of the current node coordinate axis must be the node that has not completed encoding and decoding, and the node located in the negative direction of the current node coordinate axis must be the node that has completed encoding and decoding. Therefore, For the surrounding child nodes of the first child node, it is only necessary to consider the nodes in the negative direction that have completed encoding and decoding, and do not need to consider the nodes in the positive direction whose occupancy information is unknown.
  • the Morton order only needs to consider nodes in the negative direction, there are only 3 surrounding child nodes with a distance of 1 from the first child node, and the distance from the first child node is The surrounding sub-nodes only include 3, and the distance from the first sub-node is There is only one surrounding child node of , please refer to Figure 7, the dark node in Figure 7 is the first child node, and the remaining 7 nodes are the surrounding child nodes of the first child node based on Morton order.
  • the encoding and decoding of point cloud geometry based on the Hilbert order is different.
  • the encoding and decoding order based on the Hilbert order is not fixed.
  • Even the node located in the positive direction of the current node coordinate axis may be the node that has completed encoding and decoding. , therefore, for the surrounding child nodes of the first child node, the nodes located in the positive direction of the current node coordinate axis can also be considered, so that there can be 6 surrounding child nodes with a distance of 1 from the first child node.
  • the node distance is There can be 12 surrounding child nodes of , and the distance from the first child node is There can be 8 surrounding child nodes of .
  • the surrounding child nodes of the first child node may not complete the encoding and decoding.
  • the occupancy information is unknown (in order to maintain unity with the decoding end, the encoding end will also consider the occupancy information of the surrounding sub-nodes that have not completed encoding and decoding as unknown).
  • the occupancy information of the surrounding sub-nodes is treated as the occupancy information that does not include point cloud points.
  • the occupancy information of these surrounding sub-nodes that have not completed encoding and decoding can also be treated as the occupancy information that includes point cloud points. .
  • the geometric coordinates of each child node can be converted into the corresponding Hilbert code, and the Hilbert code can be used according to the Hilbert code.
  • the size of each child node is sorted from large to small or from small to large, so as to obtain the Hilbert order corresponding to each child node.
  • There are many ways to convert geometric coordinates into Hilbert codes for example, it can be determined by looking up a table, or it can be converted by mathematical operations.
  • the space occupation code corresponding to the first node generated by the encoding end is obtained by arranging the occupation information of each sub-node corresponding to the first node in the order of Hilbert order, so that , the decoding end also needs to arrange the sub-nodes in the order of Hilbert order, so that the target sub-nodes containing point cloud points in each sub-node can be determined according to the space occupancy code.
  • the point cloud points have been arranged in the order of the Hilbert order, and there is no need to reorder the point cloud points, which greatly facilitates the follow-up based on the Hilbert order.
  • the encoding and decoding of point cloud attributes improves the overall achievability of the point cloud encoding and decoding scheme.
  • FIG. 8 is a flowchart of a point cloud decoding method provided by an embodiment of the present application. The method may include the following steps:
  • the first node may be any non-leaf node containing point cloud points obtained by tree-like division of the point cloud.
  • the decoding end since the decoding end does not know the geometric coordinates of each point cloud point like the encoding end, therefore, except for the division of the root node, the division of other nodes needs to be carried out according to the instructions of the space occupancy code.
  • which sub-nodes are target sub-nodes containing point cloud points can be determined according to the space occupancy code, so that the target sub-nodes can be further divided.
  • the target sub-nodes are already leaf nodes (the smallest space unit). Then, the geometric coordinates of the point cloud points can be determined directly according to the geometric coordinates of the target child nodes.
  • the space occupation code corresponding to the first node is obtained by decoding (at this time, the first node has been divided according to the space occupation code obtained by decoding before), since the space occupation code is the occupation information of each child node of the first node according to Hill Therefore, it is necessary to first determine the Hilbert order corresponding to each child node, arrange each child node in the order of the Hilbert order, and then compare the first node obtained by decoding Space occupancy code, which determines the target child nodes that contain point cloud points in each child node.
  • the points of each point cloud have been arranged in the order of Hilbert order, and can be used for encoding and decoding of point cloud attributes based on the Hilbert order without reordering.
  • the efficiency of encoding and decoding point cloud data is improved, and the achievability of the overall solution for encoding and decoding point cloud data based on the Hilbert sequence is improved.
  • the space occupancy code corresponding to the first node is obtained by decoding from the code stream, including:
  • the space occupancy code corresponding to the first node is obtained by decoding from the code stream.
  • context-based adaptive binary arithmetic decoding to decode the code stream to obtain the space occupancy code corresponding to the first node, including:
  • the first occupation information is the occupation information of the first child node, and the first child node is any child node in the multiple child nodes;
  • the determining the context model corresponding to the first occupancy information includes:
  • a context model corresponding to the first occupancy information is determined according to the spatial position of the first child node in the first node.
  • the determining the context model corresponding to the first occupancy information includes:
  • a context model corresponding to the first occupancy information is determined according to occupancy information of a plurality of surrounding sub-nodes whose distances from the first sub-node are a first preset value.
  • the first preset value includes one or more of the following values: 1.
  • the occupancy information of the undecoded surrounding sub-nodes is determined as not including the occupancy information of the point cloud points.
  • the multiple surrounding child nodes are divided into multiple child node groups, and the first The context model corresponding to the occupancy information, including:
  • a context model corresponding to the first occupancy information is determined according to occupancy information corresponding to each of the child node groups.
  • the determining the occupancy information corresponding to each of the sub-node groups includes:
  • any sub-node group in the plurality of sub-node groups if any surrounding sub-node in the sub-node group contains a point cloud point, it is determined that the occupancy information of the sub-node group is an occupancy including a point cloud point information;
  • the occupancy information of the sub-node group is occupancy information that does not contain point cloud points.
  • determining the context model corresponding to the first occupancy information according to the occupancy information corresponding to each of the sub-node groups includes:
  • the context model corresponding to the first occupancy information is the context model corresponding to the first index.
  • the occupancy information used to indicate that the occupancy information including the point cloud points is the first binary symbol
  • the occupancy information used to indicate that the occupancy information that does not include the point cloud points is the second binary symbol
  • the first index is a value corresponding to a binary symbol string obtained by arranging the occupancy information corresponding to each sub-node group.
  • the 6 surrounding sub-nodes whose distance from the first sub-node is 1 are divided into three sub-node groups corresponding to the X-axis direction, the Y-axis direction, and the Z-axis direction. .
  • the distance from the first child node is The 12 surrounding sub-nodes are divided into three sub-node groups corresponding to the XY plane, the XZ plane and the YZ plane.
  • the distance from the first child node is The 8 surrounding child nodes of are divided into a child node group.
  • the method further includes:
  • the occupancy information of the plurality of surrounding sub-nodes all indicate that no point cloud points are included, then determine the corresponding occupancy information of the first occupancy information according to the occupancy information of the surrounding nodes whose distance from the first node is the second preset value. context model.
  • the determining a context model corresponding to the first occupancy information according to occupancy information of a neighboring node whose distance from the first node is a second preset value includes:
  • the context model corresponding to the first occupancy information is the context model corresponding to the second index.
  • the second index is a value corresponding to a binary symbol string obtained by arranging the occupancy information of each of the surrounding nodes.
  • the determining the Hilbert order corresponding to the multiple child nodes by using the geometric coordinates of the multiple child nodes includes:
  • the space occupation code corresponding to the first node generated by the encoding end is obtained by arranging the occupation information of each sub-node corresponding to the first node in the order of Hilbert order, so that , the decoding end also needs to arrange the sub-nodes in the order of Hilbert order, so that the target sub-nodes containing point cloud points in each sub-node can be determined according to the space occupancy code.
  • the point cloud points have been arranged in the order of the Hilbert order, and there is no need to reorder the point cloud points, which greatly facilitates the follow-up based on the Hilbert order.
  • the encoding and decoding of point cloud attributes improves the overall achievability of the point cloud encoding and decoding scheme.
  • FIG. 9 is a schematic structural diagram of a point cloud encoding apparatus provided by an embodiment of the present application.
  • the apparatus may include: a processor 910 and a memory 920 storing a computer program, the processor implements the following steps when executing the computer program:
  • the first node is divided into a plurality of sub-nodes, and the first node is any non-leaf node containing point cloud points obtained by tree-like division of the point cloud;
  • the space occupation code is encoded.
  • the processor when encoding the space occupation code, is configured to use context-based adaptive binary arithmetic coding to encode the space occupation code.
  • the processor when the processor encodes the space occupancy code by using context-based adaptive binary arithmetic coding, the processor is used to determine a context model corresponding to the first occupancy information, where the first occupancy information is the first child Occupancy information of a node, the first child node is any child node of the plurality of child nodes; arithmetic coding is performed on the first occupancy information according to the context model.
  • the processor when determining the context model corresponding to the first occupancy information, is configured to determine, according to the spatial position of the first child node in the first node, the context model corresponding to the first occupancy information .
  • the processor when determining the context model corresponding to the first occupancy information, is configured to: The context model corresponding to the first occupancy information is described.
  • the first preset value includes one or more of the following values: 1.
  • the occupancy information of the uncoded surrounding sub-nodes is determined as not including the occupancy information of the point cloud points.
  • the multiple surrounding child nodes are divided into multiple child node groups, and the processor determines the number of surrounding child nodes according to the occupancy information of the surrounding child nodes whose distance from the first child node is a first preset value.
  • the context model corresponding to the first occupancy information is used, the occupancy information corresponding to each of the sub-node groups is determined; and the context model corresponding to the first occupancy information is determined according to the occupancy information corresponding to each of the sub-node groups.
  • the processor determines the occupancy information corresponding to each of the sub-node groups, for any sub-node group in the multiple sub-node groups, if any surrounding in the sub-node group is If the child node contains point cloud points, it is determined that the occupancy information of the child node group is the occupancy information containing point cloud points; if all surrounding child nodes in the child node group do not contain point cloud points, then determine that the child node group The occupancy information of the node group is the occupancy information that does not contain point cloud points.
  • the processor when determining the context model corresponding to the first occupancy information according to the occupancy information corresponding to each of the sub-node groups, is configured to determine the first occupancy information according to the occupancy information corresponding to each of the sub-node groups. index; determining that the context model corresponding to the first occupancy information is the context model corresponding to the first index.
  • the occupancy information used to indicate that the occupancy information including the point cloud points is the first binary symbol
  • the occupancy information used to indicate that the occupancy information that does not include the point cloud points is the second binary symbol
  • the first index is a numerical value corresponding to a binary symbol string obtained by arranging the occupancy information corresponding to each sub-node group.
  • the 6 surrounding sub-nodes whose distance from the first sub-node is 1 are divided into three sub-node groups corresponding to the X-axis direction, the Y-axis direction, and the Z-axis direction. .
  • the distance from the first child node is The 12 surrounding sub-nodes are divided into three sub-node groups corresponding to the XY plane, the XZ plane and the YZ plane.
  • the distance from the first child node is The 8 surrounding child nodes of are divided into a child node group.
  • the processor is further configured to, if the occupancy information of the plurality of surrounding sub-nodes all indicate that no point cloud points are included, according to the surrounding nodes whose distance from the first node is the second preset value.
  • the occupancy information of the first occupancy information determines the context model corresponding to the first occupancy information.
  • the processor when determining the context model corresponding to the first occupancy information according to the occupancy information of the surrounding nodes whose distance from the first node is the second preset value, is configured to: The occupancy information of the node determines a second index; and the context model corresponding to the first occupancy information is determined to be the context model corresponding to the second index.
  • the second index is a value corresponding to a binary symbol string obtained by arranging the occupancy information of each of the surrounding nodes.
  • the processor when using the geometric coordinates of the multiple child nodes to determine the Hilbert order corresponding to the multiple child nodes, the processor is used to convert the geometric coordinates of each of the child nodes into corresponding Hilbert sequences. Bert code; sort each of the child nodes according to the Hilbert code to obtain the Hilbert order.
  • the division order of the first node and other nodes to be divided belonging to the same layer in the tree structure is determined according to the Hilbert order corresponding to each node in the same layer.
  • the space occupation code corresponding to the first node generated by the encoding end is obtained by arranging the occupation information of each sub-node corresponding to the first node in the order of Hilbert order, thus, The decoding end also needs to arrange the sub-nodes in the order of Hilbert order, so that the target sub-nodes including the point cloud points in each sub-node can be determined according to the space occupation code.
  • the point cloud points after completing the encoding and decoding of the geometric data of the point cloud, the point cloud points have been arranged in the order of the Hilbert order, and there is no need to reorder the point cloud points, which greatly facilitates the follow-up based on the Hilbert order.
  • the encoding and decoding of point cloud attributes improves the overall achievability of the point cloud encoding and decoding scheme.
  • FIG. 10 is a schematic structural diagram of a point cloud decoding apparatus provided by an embodiment of the present application.
  • the apparatus may include: a processor 1010 and a memory 1020 storing a computer program, the processor implements the following steps when executing the computer program:
  • the first node is divided into a plurality of sub-nodes, and the first node is any non-leaf node containing point cloud points obtained by tree-like division of the point cloud;
  • the target child node is a non-leaf node, perform the tree division on the target child node.
  • the processor is further configured to, if the target child node is a leaf node, determine the geometric coordinates of the target child node as the geometric coordinates of the point cloud point.
  • the processor when the processor decodes the code stream to obtain the space occupancy code corresponding to the first node, it is used to decode the code stream to obtain the corresponding space occupancy code of the first node by using context-based adaptive binary arithmetic decoding. space occupancy code.
  • the processor is used to determine the context model corresponding to the first occupancy information when decoding the space occupancy code corresponding to the first node from the code stream by using context-based adaptive binary arithmetic decoding, where
  • the first occupancy information is the occupancy information of the first child node, and the first child node is any child node in the plurality of child nodes;
  • the processor when determining the context model corresponding to the first occupancy information, is configured to determine, according to the spatial position of the first child node in the first node, the context model corresponding to the first occupancy information .
  • the processor when determining the context model corresponding to the first occupancy information, is configured to: The context model corresponding to the first occupancy information is described.
  • the first preset value includes one or more of the following values: 1.
  • the occupancy information of the undecoded surrounding sub-nodes is determined as not including the occupancy information of the point cloud points.
  • the multiple surrounding child nodes are divided into multiple child node groups, and the processor determines the number of surrounding child nodes according to the occupancy information of the surrounding child nodes whose distance from the first child node is a first preset value.
  • the context model corresponding to the first occupancy information is used, the occupancy information corresponding to each of the sub-node groups is determined; and the context model corresponding to the first occupancy information is determined according to the occupancy information corresponding to each of the sub-node groups.
  • the processor determines the occupancy information corresponding to each of the sub-node groups, for any sub-node group in the multiple sub-node groups, if any surrounding in the sub-node group is If the child node contains point cloud points, it is determined that the occupancy information of the child node group is the occupancy information that includes point cloud points; if all surrounding child nodes in the child node group do not contain point cloud points, it is determined that the child node group contains point cloud points.
  • the occupancy information of the node group is the occupancy information that does not contain point cloud points.
  • the processor when determining the context model corresponding to the first occupancy information according to the occupancy information corresponding to each of the sub-node groups, is configured to determine the first occupancy information according to the occupancy information corresponding to each of the sub-node groups. index; determining that the context model corresponding to the first occupancy information is the context model corresponding to the first index.
  • the occupancy information used to indicate that the occupancy information including the point cloud points is the first binary symbol
  • the occupancy information used to indicate that the occupancy information that does not include the point cloud points is the second binary symbol
  • the first index is a value corresponding to a binary symbol string obtained by arranging the occupancy information corresponding to each sub-node group.
  • the 6 surrounding sub-nodes whose distance from the first sub-node is 1 are divided into three sub-node groups corresponding to the X-axis direction, the Y-axis direction, and the Z-axis direction. .
  • the distance from the first child node is The 12 surrounding sub-nodes are divided into three sub-node groups corresponding to the XY plane, the XZ plane and the YZ plane.
  • the distance from the first child node is The 8 surrounding child nodes of are divided into a child node group.
  • the processor is further configured to, if the occupancy information of the plurality of surrounding sub-nodes all indicate that no point cloud points are included, according to the surrounding nodes whose distance from the first node is the second preset value.
  • the occupancy information of the first occupancy information determines the context model corresponding to the first occupancy information.
  • the processor when determining the context model corresponding to the first occupancy information according to the occupancy information of the surrounding nodes whose distance from the first node is the second preset value, is configured to: The occupancy information of the node determines a second index; and the context model corresponding to the first occupancy information is determined to be the context model corresponding to the second index.
  • the second index is a value corresponding to a binary symbol string obtained by arranging the occupancy information of each of the surrounding nodes.
  • the processor when using the geometric coordinates of the multiple child nodes to determine the Hilbert order corresponding to the multiple child nodes, the processor is used to convert the geometric coordinates of each of the child nodes into corresponding Hilbert sequences. Bert code; sort each of the child nodes according to the Hilbert code to obtain the Hilbert order.
  • the space occupancy code corresponding to the first node generated by the encoding end is obtained by arranging the occupancy information of each child node corresponding to the first node in the order of Hilbert order, thus, The decoding end also needs to arrange the sub-nodes in the order of Hilbert order, so that the target sub-nodes including the point cloud points in each sub-node can be determined according to the space occupation code.
  • the point cloud points after completing the encoding and decoding of the geometric data of the point cloud, the point cloud points have been arranged in the order of the Hilbert order, and there is no need to reorder the point cloud points, which greatly facilitates the follow-up based on the Hilbert order.
  • the encoding and decoding of point cloud attributes improves the overall achievability of the point cloud encoding and decoding scheme.
  • Embodiments of the present application further provide a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, any of the point cloud encoding methods provided by the embodiments of the present application is implemented.
  • Embodiments of the present application further provide a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, any of the point cloud decoding methods provided by the embodiments of the present application is implemented.
  • Embodiments of the present application may take the form of a computer program product implemented on one or more storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having program code embodied therein.
  • Computer-usable storage media includes permanent and non-permanent, removable and non-removable media, and storage of information can be accomplished by any method or technology.
  • Information may be computer readable instructions, data structures, modules of programs, or other data.
  • Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), Flash Memory or other memory technology, Compact Disc Read Only Memory (CD-ROM), Digital Versatile Disc (DVD) or other optical storage, Magnetic tape cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission medium that can be used to store information that can be accessed by a computing device.
  • PRAM phase-change memory
  • SRAM static random access memory
  • DRAM dynamic random access memory
  • RAM random access memory
  • ROM read only memory
  • EEPROM Electrically Erasable Programmable Read Only Memory
  • Flash Memory or other memory technology
  • CD-ROM Compact Disc Read Only Memory
  • CD-ROM Compact Disc Read Only Memory
  • DVD Digital Versatile Disc
  • Magnetic tape cassettes magnetic tape magnetic disk storage or other magnetic storage devices or any other non-

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)
  • Compression, Expansion, Code Conversion, And Decoders (AREA)

Abstract

L'invention concerne un procédé de codage de nuage de points consistant à : diviser un premier nœud en une pluralité de sous-nœuds (202), le premier nœud étant un nœud sans feuille quelconque obtenu en effectuant une répartition d'arborescence sur un nuage de points et comprenant un point de nuage de points ; déterminer, à l'aide des coordonnées géométriques de la pluralité de sous-nœuds, une séquence de Hilbert correspondant à la pluralité de sous-nœuds (204) ; agencer les informations d'occupation de la pluralité de sous-nœuds selon la séquence de Hilbert de façon à obtenir un code d'occupation d'espace correspondant au premier nœud (206), les informations d'occupation servant à indiquer si les sous-nœuds comprennent des points de nuage de points ; et coder le code d'occupation d'espace (208). Le procédé permet de résoudre le problème technique de la nécessité de réagencer des points de nuage de points, qui sont reconstruits après un codage et un décodage géométriques, avant d'effectuer un codage et un décodage d'attributs sur un nuage de points.
PCT/CN2020/134355 2020-12-07 2020-12-07 Procédé et appareil de codage de nuage de points, procédé et appareil de décodage de nuage de points, et support de stockage lisible par ordinateur WO2022120542A1 (fr)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/CN2020/134355 WO2022120542A1 (fr) 2020-12-07 2020-12-07 Procédé et appareil de codage de nuage de points, procédé et appareil de décodage de nuage de points, et support de stockage lisible par ordinateur
CN202080081330.1A CN114885617A (zh) 2020-12-07 2020-12-07 点云编码和解码方法、装置及计算机可读存储介质

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2020/134355 WO2022120542A1 (fr) 2020-12-07 2020-12-07 Procédé et appareil de codage de nuage de points, procédé et appareil de décodage de nuage de points, et support de stockage lisible par ordinateur

Publications (1)

Publication Number Publication Date
WO2022120542A1 true WO2022120542A1 (fr) 2022-06-16

Family

ID=81972813

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/134355 WO2022120542A1 (fr) 2020-12-07 2020-12-07 Procédé et appareil de codage de nuage de points, procédé et appareil de décodage de nuage de points, et support de stockage lisible par ordinateur

Country Status (2)

Country Link
CN (1) CN114885617A (fr)
WO (1) WO2022120542A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2024074123A1 (fr) * 2022-10-04 2024-04-11 Douyin Vision Co., Ltd. Procédé, appareil et support de codage en nuage de points

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103617162A (zh) * 2013-10-14 2014-03-05 南京邮电大学 一种对等云平台上构建希尔伯特r树索引的方法
CN110915219A (zh) * 2017-07-13 2020-03-24 交互数字Vc控股公司 用于对几何形状由基于八叉树的结构表示的有色点云的颜色进行编码/解码的方法和装置
WO2020123469A1 (fr) * 2018-12-11 2020-06-18 Futurewei Technologies, Inc. Encodage d'attribut d'arbre hiérarchique par points médians dans un codage en nuage de points

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103617162A (zh) * 2013-10-14 2014-03-05 南京邮电大学 一种对等云平台上构建希尔伯特r树索引的方法
CN110915219A (zh) * 2017-07-13 2020-03-24 交互数字Vc控股公司 用于对几何形状由基于八叉树的结构表示的有色点云的颜色进行编码/解码的方法和装置
WO2020123469A1 (fr) * 2018-12-11 2020-06-18 Futurewei Technologies, Inc. Encodage d'attribut d'arbre hiérarchique par points médians dans un codage en nuage de points

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
LIU, HUI ET AL.: "Development of Encoding and Decoding Algorithms for High Dimensional Hilbert Curves", JOURNAL ON NUMERICAL METHODS AND COMPUTER APPLICATIONS, vol. 36, no. 1, 31 March 2015 (2015-03-31), pages 42 - 58, XP055942127 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2024074123A1 (fr) * 2022-10-04 2024-04-11 Douyin Vision Co., Ltd. Procédé, appareil et support de codage en nuage de points

Also Published As

Publication number Publication date
CN114885617A (zh) 2022-08-09

Similar Documents

Publication Publication Date Title
JP5456903B2 (ja) メッシュ・モデルを符号化する方法及び装置、符号化されたメッシュ・モデル、並びに、メッシュ・モデルを復号化する方法及び装置
JP5955378B2 (ja) エンコード方法及びデコード方法
US9721320B2 (en) Fully parallel in-place construction of 3D acceleration structures and bounding volume hierarchies in a graphics processing unit
JP7330306B2 (ja) 変換方法、逆変換方法、エンコーダ、デコーダ及び記憶媒体
Villanueva et al. SSVDAGs: Symmetry-aware sparse voxel DAGs
US10019649B2 (en) Point cloud simplification
WO2022121649A1 (fr) Procédé de codage et de décodage de données de nuage de points, procédé et appareil de traitement de données de nuage de points, dispositif électronique, produit-programme d'ordinateur et support de stockage lisible par ordinateur
Kontkanen et al. Coherent out‐of‐core point‐based global illumination
EP4088261A1 (fr) Détermination de contexte pour un mode planaire dans un codage en nuage de points fondé sur un octree
CN113094463A (zh) 一种非结构化点云存储方法、装置、设备及介质
WO2022120542A1 (fr) Procédé et appareil de codage de nuage de points, procédé et appareil de décodage de nuage de points, et support de stockage lisible par ordinateur
CN113468286A (zh) 一种基于三角面片个数划分的三维金字塔构建方法
CN110825830B (zh) 一种网格空间的数据检索方法
CN116258840B (zh) 层级细节表示树的生成方法、装置、设备及存储介质
Pinto et al. Improved queryable representations of rasters
CN117009411A (zh) 一种基于点云数据网格化空间存储与索引方法、装置及计算机可读存储介质
CN114116925A (zh) 一种时空数据的查询方法及相关装置
CN113032405A (zh) 时空数据管理方法、系统、主机及计算机可读存储介质
Jakob et al. A parallel approach to compression and decompression of triangle meshes using the GPU
CN113849495A (zh) 一种点云动态哈希划分方法及设备
Gang et al. Research on spatial index structure of massive point clouds based on hybrid tree
Molenaar et al. Editing Compressed High‐resolution Voxel Scenes with Attributes
WO2022155929A1 (fr) Procédé et dispositif de décodage de géométrie de nuage de points et support de stockage lisible par ordinateur
CN117744185B (zh) 几何模型的粒子生成方法、装置、电子设备及存储介质
Dolonius Sparse Voxel DAGs for Shadows and for Geometry with Colors

Legal Events

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

Ref document number: 20964489

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20964489

Country of ref document: EP

Kind code of ref document: A1