WO2021043136A1 - Point cloud projection transmission method, and reception processing method and system - Google Patents
Point cloud projection transmission method, and reception processing method and system Download PDFInfo
- Publication number
- WO2021043136A1 WO2021043136A1 PCT/CN2020/112857 CN2020112857W WO2021043136A1 WO 2021043136 A1 WO2021043136 A1 WO 2021043136A1 CN 2020112857 W CN2020112857 W CN 2020112857W WO 2021043136 A1 WO2021043136 A1 WO 2021043136A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- point cloud
- projection
- segment
- user
- perspective
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 171
- 230000005540 biological transmission Effects 0.000 title claims abstract description 166
- 238000003672 processing method Methods 0.000 title claims abstract description 26
- 239000013598 vector Substances 0.000 claims abstract description 365
- 239000012634 fragment Substances 0.000 claims abstract description 226
- 238000012545 processing Methods 0.000 claims description 67
- 230000008569 process Effects 0.000 claims description 45
- 238000007906 compression Methods 0.000 claims description 44
- 230000006835 compression Effects 0.000 claims description 44
- 230000008859 change Effects 0.000 claims description 25
- 238000012986 modification Methods 0.000 claims description 19
- 230000004048 modification Effects 0.000 claims description 19
- 238000013507 mapping Methods 0.000 claims description 16
- 238000005457 optimization Methods 0.000 abstract description 26
- 230000003044 adaptive effect Effects 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 25
- 238000005516 engineering process Methods 0.000 description 14
- 238000003860 storage Methods 0.000 description 11
- 230000000694 effects Effects 0.000 description 10
- 238000004364 calculation method Methods 0.000 description 9
- 238000004422 calculation algorithm Methods 0.000 description 8
- 230000009467 reduction Effects 0.000 description 8
- 230000006870 function Effects 0.000 description 5
- 230000003287 optical effect Effects 0.000 description 5
- 230000008447 perception Effects 0.000 description 4
- 238000010561 standard procedure Methods 0.000 description 4
- 238000004590 computer program Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 238000011946 reduction process Methods 0.000 description 3
- 238000009877 rendering Methods 0.000 description 3
- 230000006837 decompression Effects 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000013144 data compression Methods 0.000 description 1
- 238000013467 fragmentation Methods 0.000 description 1
- 238000006062 fragmentation reaction Methods 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000005549 size reduction Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N13/00—Stereoscopic video systems; Multi-view video systems; Details thereof
- H04N13/10—Processing, recording or transmission of stereoscopic or multi-view image signals
- H04N13/106—Processing image signals
- H04N13/111—Transformation of image signals corresponding to virtual viewpoints, e.g. spatial image interpolation
- H04N13/117—Transformation of image signals corresponding to virtual viewpoints, e.g. spatial image interpolation the virtual viewpoint locations being selected by the viewers or determined by viewer tracking
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N13/00—Stereoscopic video systems; Multi-view video systems; Details thereof
- H04N13/10—Processing, recording or transmission of stereoscopic or multi-view image signals
- H04N13/106—Processing image signals
- H04N13/122—Improving the 3D impression of stereoscopic images by modifying image signal contents, e.g. by filtering or adding monoscopic depth cues
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N13/00—Stereoscopic video systems; Multi-view video systems; Details thereof
- H04N13/10—Processing, recording or transmission of stereoscopic or multi-view image signals
- H04N13/106—Processing image signals
- H04N13/139—Format conversion, e.g. of frame-rate or size
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N13/00—Stereoscopic video systems; Multi-view video systems; Details thereof
- H04N13/10—Processing, recording or transmission of stereoscopic or multi-view image signals
- H04N13/106—Processing image signals
- H04N13/161—Encoding, multiplexing or demultiplexing different image signal components
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N13/00—Stereoscopic video systems; Multi-view video systems; Details thereof
- H04N13/10—Processing, recording or transmission of stereoscopic or multi-view image signals
- H04N13/194—Transmission of image signals
-
- 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
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/597—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding specially adapted for multi-view video sequence encoding
Definitions
- the invention belongs to the field of point clouds, and in particular relates to a point cloud projection transmission method, reception processing method and system based on a user's perspective.
- Three-dimensional point cloud data is a new type of image media data, which can provide users with more accurate and vivid image information, and bring higher fidelity and more interactive communication experience.
- Point cloud data is the surface information of the object obtained by scanning, including three-dimensional coordinate data, depth information, color information, and so on.
- the maturity of 3D scanning technology and the perfection of the system architecture have improved the accuracy of point cloud data, while also greatly increasing the amount of point cloud data.
- point cloud data compression algorithms for static and dynamic point cloud data, and different types of point cloud data such as machine perception and human eye perception.
- a typical point cloud compression algorithm is to convert 3D point cloud data into 2D image data, and then perform data processing, one of which is video-based point cloud compression (Video- based Point Cloud Compression, VPCC) algorithm.
- Video- based Point Cloud Compression Video- based Point Cloud Compression
- This compression method first projects a 3D point cloud onto a 2D plane to obtain occupancy map information, geometric information, attribute information, and auxiliary information.
- the attribute information usually includes texture information and color information. Therefore, the compressed information is usually divided into four categories. Data is transferred. Through this method, the overall compressed transmission of the point cloud sequence is realized.
- a typical compression scheme applied to point cloud coding and transmission is to convert three-dimensional point cloud data into two-dimensional images for processing through segmented projection.
- This solution uses the existing video coding tools to encode and transmit the point cloud, in which it is necessary to sequentially carry out the division of the point cloud, the subdivision, segment projection, blank filling and other steps to achieve the segment projection of the point cloud. Then, the two-dimensional data obtained by projection can be encoded and transmitted with the help of video encoding technologies such as HEVC. Through this method, the overall compressed transmission of the point cloud sequence is realized.
- Patent document CN104778691B (application number: 201510160208.4) discloses a simplified processing method for three-dimensional point cloud data.
- the invention mainly solves the technical problems of poor retention of three-dimensional object surface features, poor reconstruction effect, and large amount of calculation existing in the existing method for streamlining three-dimensional point clouds.
- the invention uses the information in the color image synchronized by the Kinect camera to streamline the three-dimensional point cloud when collecting the point cloud data.
- the method combines the color map and the point cloud for streamlining, which can avoid some curvature of the object surface caused by the collection and streamlining process There is no significant change but the point cloud data of the characteristic information is lost; you can set the corresponding reduction rate according to the gray information of the object itself for reduction; you can also reserve the characteristic points that you think are more important according to subjective factors, and selectively simplify Point cloud.
- the user cannot observe the entire content of a point cloud object.
- the user cannot see the media content on the back of the point cloud.
- the process of cloud processing does not differentiate users' attention to differences in various aspects.
- the existing point cloud compression scheme implements point cloud transmission. Since the point cloud fragments that can be observed and the point cloud fragments that cannot be observed are compressed with the same accuracy and provide the same definition of presentation, this will Unnecessary demand for transmission bandwidth, etc. will be put forward. For this unnecessary amount of information caused by the occlusion of the point cloud, the current overall compression transmission scheme for the point cloud sequence does not propose a corresponding solution. How to make distinctions based on the degree of attention of the users' different perspectives, and to reduce this information redundancy without affecting the quality of observations, and adapt it to actual user application scenarios, are key issues that need to be resolved urgently.
- the purpose of the present invention is to provide a point cloud projection transmission method, reception processing method and system, which introduces the user's perspective to improve the quality of point cloud projection processing, and optionally, projection is performed on the side with high attention that has a greater impact on the user's perception
- the vector is modified and perfected, or combined with optimized point cloud compression algorithm to achieve.
- the present invention provides a point cloud projection transmission method based on the user's perspective, which has such features, including: preliminary division of point cloud objects on a standard vector to obtain the point cloud segment to be processed; and determination based on the user's perspective
- the projection transmission scheme is to process the point cloud segment to be processed.
- the point cloud projection transmission method may also have the feature that the closest standard normal vector is taken as the classification category of the midpoint of the point cloud object; based on six standards
- the normal vector divides the midpoint of the point cloud object into multiple first-level point cloud fragments; in the same first-level point cloud fragment, the points with the same normal vector and the distance less than the predetermined value are then divided into second-level point cloud fragments as to be processed Cloud fragment.
- the point cloud projection transmission method may also have such a feature, wherein the projection transmission scheme includes: further dividing the to-be-processed point cloud segment to obtain points corresponding to the perspective category Cloud fragments, including the point cloud fragments that need to be observed and the point cloud fragments that do not need to be observed; the point cloud fragments that need to be observed are assigned priority resources for encoding transmission, and the point cloud fragments that do not need to be observed are compared Prioritize the allocation of secondary resources for coding transmission.
- the point cloud projection transmission method based on the user's perspective provided by the present invention, it may also have such a feature, wherein the projection transmission scheme includes any one or more of the following: no point cloud modification Projection plan, improve the point cloud projection plan, modify the standard normal vector plan according to the user's perspective.
- the solution of not modifying the point cloud projection method includes: placing the point cloud object in the standard vector
- the above preliminary division is to associate the parameter list with the standard vector as the projection vector to obtain the point cloud segment to be processed.
- the point cloud segment to be processed is divided into the point cloud segment that needs to be observed and the point cloud segment that needs to be observed.
- the point cloud projection transmission method may also have the feature of setting corresponding points for the point cloud fragments that need to be observed and the point cloud fragments that do not need to be observed.
- the encoding and decoding parameters of each point cloud segment are recorded as additional information for encoding and transmission.
- the improved point cloud projection solution includes: further dividing the to-be-processed point cloud fragments to obtain corresponding
- the point cloud fragments of the perspective category include the point cloud fragments that need to be observed and the point cloud fragments that do not need to be observed; the point cloud fragments that need to be observed and the point cloud fragments that do not need to be observed are projected onto the projection plane, according to The size of the projection plane needs to be changed to obtain the need to observe pictures and the need not to observe pictures.
- the point cloud projection transmission method may also have such characteristics, and further include: as the first improved point cloud projection solution, the point cloud segment and the Under the premise that the point cloud fragments that do not need to be observed are misaligned, the pictures that do not need to be observed are merged into the pictures that need to be observed, and a projected picture is obtained, which is compressed and encoded.
- the point cloud projection transmission method may also have the feature that a projected picture obtained forms a set of video sequences, and the point cloud that does not need to be observed is recorded
- the index parameter index corresponding to the segment and the scale of the size change are used as newly added additional information for encoding transmission.
- the point cloud projection transmission method may also have such characteristics, and further include: as a second improved point cloud projection solution, there will be a need to observe the picture and the need not to observe the picture. Compress and encode separately.
- the point cloud projection transmission method may also have the feature that the projected images that need to be observed and the images that do not need to be observed respectively form two sets of video sequences, and the two sets are encoded.
- Video sequence record the change ratio of the video sequence corresponding to the point cloud segment that does not need to be observed, encode and transmit as new additional information, and further record the position of each group of video sequences in the coded stream.
- the third improved point cloud projection solution includes: the point cloud segment to be processed according to six standard methods The vectors are grouped and projected onto six projection planes, and the size of the six projection planes is changed according to the user's perspective based on the user's perspective, and compression coding is performed respectively.
- the point cloud segment corresponding to each standard normal vector can be respectively projected to each group of video sequences, and six groups are encoded.
- Video sequence record the change ratio of the video sequence corresponding to the point cloud segment that does not need to be observed, encode and transmit as new additional information, and further record the position of each group of video sequences in the coded stream.
- the modification of the standard normal vector scheme according to the user's perspective includes: determining the perspective according to the user's observation perspective Vector; rotate the view vector with one of the standard normal vectors, and then perform the same rotation mapping processing on the other standard normal vectors, and determine the projection direction after the rotation mapping on the premise that the overall modification amount is less than a predetermined value.
- the point cloud projection transmission method based on the user's perspective may also have the feature, wherein the step of reducing the overall modification amount to be less than a predetermined value includes: the perspective vector corresponds to the standard normal vector , And after the rotation, the sum of the dot products of the six standard normal vectors after the rotation mapping including the viewing angle vector and the corresponding original six standard normal vectors takes the maximum value or a larger value within a predetermined range.
- the point cloud projection transmission method may also have the feature that the standard normal vector after the rotation mapping is adjusted will be encoded as the newly added additional information transmission.
- the point cloud projection transmission method based on the user's perspective provided by the present invention, it may also have such a feature, wherein the projection transmission scheme includes: the point cloud segment to be processed is distinguished from the main segment and the point cloud segment according to the user's perspective. Multiple edge segments; considering the spatial connection between the edge segment and the main segment, the projection vector of the edge segment is maintained or changed based on the user's perspective; subsequent point cloud processing is performed after projection.
- the judgment basis considering the spatial connection between the edge segment and the main segment includes: The user’s perspective determines the principal vector in the standard normal vector; the principal segment and the edge segment are respectively projected along the principal vector direction to obtain the principal two-dimensional projection map and the edge two-dimensional projection map; based on each edge two-dimensional projection map, and each The main two-dimensional projection map is calculated to obtain the intersection ratio; multiple intersection ratios are compared with the preset threshold value as the basis for judging the spatial connection.
- the point cloud projection transmission method may also have the feature that when multiple intersection ratios are greater than a preset threshold, the edge The projection vector of the segment is changed to the main vector; if none of the multiple intersection ratios is greater than the preset threshold, the projection vector of the edge segment is maintained.
- the preset threshold value depends on the surface characteristics of the point cloud object to distinguish the object point cloud object.
- the determination of the threshold value of the surface characteristics of the character point cloud object and the surface characteristics of the character point cloud object can be further adjusted and determined according to the experimental results.
- the point cloud projection transmission method may also have the feature, wherein, according to the position where the user views the point cloud object, a point cloud is used to point the user to the point cloud.
- the vector of the observation point represents the viewing angle of the user, which is called the viewing angle vector.
- the distance between the viewing angle vector and the six standard normal vectors is calculated, and the standard normal vector with the closest distance to the viewing angle vector is called the principal vector under the viewing angle.
- the point cloud projection transmission method based on the user's perspective may also have the feature, wherein, in addition to distinguishing the main segment and multiple edge segments of the point cloud object according to the user's perspective, Including distinguishing irrelevant fragments, among which, the standard normal vector closest to the user's perspective distance is determined as the main vector; among multiple standard normal vectors, the point cloud fragments to be processed that are the same as the main vector are determined as the main fragment; The to-be-processed point cloud segment opposite to the main vector in the standard normal vector is determined as an irrelevant segment; the to-be-processed point cloud segment associated with the main vector in the multiple standard normal vectors is determined as an edge segment.
- the point cloud projection transmission method may also have the feature that, in the subsequent point cloud processing after projection, the projection relationship identification field is used to identify that the point cloud The projection vector of the point in the object after processing and adjustment, including the index value of the point cloud segment and the corresponding projection vector value.
- the present invention also provides a point cloud receiving and processing method based on the user's perspective, which is characterized in that the codec parameters determined by the user's perspective are used to obtain a video sequence from the received code stream, and the video sequence is processed to obtain a point cloud. Segment, and decode and reconstruct each point cloud segment to obtain a point cloud object.
- the point cloud receiving and processing method may also have the feature, in which the process of encoding, decoding and subsequent reconstruction of each point cloud segment according to the encoding and decoding parameters, Contains any one or more of the following: the encoding and decoding parameters are set at the sending end for each point cloud segment, distinguishing the point cloud segment that needs to be observed and the point cloud segment that does not need to be observed, and performing subsequent point cloud object reconstruction; or According to the received parameter information and size change information, the point cloud segment is restored to its original size, and then the subsequent point cloud object reconstruction is performed; or each video sequence is restored according to the position of each video sequence in the code stream.
- the present invention also provides a point cloud projection transmission method based on the user's perspective, which is characterized in that: a segment generation module is used to preliminarily divide the point cloud object on the standard vector to obtain the point cloud segment to be processed; projection coding processing The module determines the projection transmission scheme based on the user's perspective, and processes the point cloud segment to be processed.
- the present invention also provides a point cloud receiving processing system based on the user's perspective, which is characterized in that: the inverse processing module uses the codec parameters determined from the user's perspective to obtain a video sequence from the received code stream, and to obtain a video sequence for the video sequence.
- the point cloud segment is obtained by processing, and each point cloud segment is decoded and reconstructed to obtain a point cloud object.
- the present invention also provides a point cloud projection transmission method based on the user's perspective, which is characterized in that: the point cloud object is preliminarily divided on the standard vector to obtain the point cloud segment to be processed; the point cloud segment to be processed is further divided to obtain
- the point cloud segment corresponding to the perspective category includes the point cloud segment that needs to be observed and the point cloud segment that does not need to be observed; priority resources are assigned to the point cloud segment that needs to be observed for encoding and transmission, and for the perspective that does not need to be observed Point cloud fragments are compared with priority allocation of secondary resources for encoding transmission
- the present invention also provides a point cloud projection transmission method based on the user's perspective, which may also have the following characteristics: including: the point cloud segment to be processed is distinguished from the main segment and multiple edge segments according to the user's perspective; The spatial connection between the fragment and the main fragment is to maintain the projection vector of the edge fragment or change it based on the user's perspective; after the projection, the subsequent point cloud processing is performed.
- the present invention also provides a point cloud projection transmission method based on the user's perspective, which may also have such features: including: preliminary division of the point cloud objects on the standard vector to obtain the to-be-processed point cloud segment, which is distinguished according to the user's perspective Generate the main segment and multiple edge segments; consider the spatial connection between the edge segment and the main segment, maintain the projection vector of the edge segment or change it based on the user's perspective; use the adjusted and changed main segment as the point to be observed Cloud fragments, adjusted and changed edge fragments and irrelevant fragments as point cloud fragments that do not need to be observed; and assign priority resources to the point cloud fragments that need to be observed for encoding transmission, and point cloud fragments that do not need to be observed. Compared with priority allocation of secondary resources for coding transmission.
- the point cloud object is preliminarily divided on the standard vector to obtain the point cloud segment to be processed
- the projection transmission scheme is determined based on the user's perspective
- the projection is adaptively adjusted according to the user's perspective
- the transmission scheme is optimized and adjusted in the direction of the user's attention to the point cloud object to improve the transmission quality.
- FIG. 1 is a process diagram of processing point cloud fragments from different perspectives without modifying the point cloud projection scheme in the first embodiment
- Figure 2-1 is a process diagram of processing point cloud fragments from the first perspective under the improved point cloud projection scheme in the first embodiment
- Fig. 2-2 is a process diagram of processing point cloud fragments from a second perspective under the improved point cloud projection scheme in the first embodiment
- Figure 2-3 is a process diagram of processing point cloud fragments from the third perspective under the improved point cloud projection scheme in the first embodiment
- FIG. 3 is an example diagram of different aspects that need to be mainly observed under different user perspectives in the first embodiment
- FIG. 4 is a schematic block diagram of the system of the sending end in the first embodiment
- FIG. 5 is a three-dimensional schematic diagram of a three-dimensional edge box of a point cloud object in the second embodiment
- FIG. 6 is a diagram of the projection adjustment process of the point cloud segment under the projection optimization adjustment scheme in the second embodiment
- FIG. 7 is a corresponding table of the relationship between normal vectors and principal vectors of different point cloud segment types in the second embodiment
- 8-1 is a schematic diagram of the relative positions of the three-dimensional edge box of the main segment and the three-dimensional edge box of the edge segment used for projection adjustment according to the main vector direction in the second embodiment;
- Figure 8-2 is a schematic diagram of a two-dimensional projection obtained after the three-dimensional edge boxes of the main segment and the edge segment in Figure 11-1 are projected onto a two-dimensional plane;
- FIG. 9 is a schematic flowchart of a point cloud projection transmission method at the sending end in the second embodiment.
- FIG. 10 is a schematic flowchart of a point cloud projection receiving method at the receiving end in the second embodiment.
- Fig. 11 is a schematic block diagram of a system at the sending end in the second embodiment.
- the present invention provides a point cloud projection transmission method based on a user's perspective, which includes: preliminarily dividing point cloud objects on a standard vector to obtain a point cloud segment to be processed; and determining a projection transmission scheme based on the user's perspective, and the point cloud segment to be processed To process.
- the present invention specifically describes different embodiments of the projection transmission scheme in the following.
- a point cloud projection transmission method based on a user's perspective provided according to the present invention includes:
- Point cloud segment determination step determine the corresponding point cloud segment according to the observation perspective, and obtain a series of point cloud segments classified into the corresponding perspective category;
- Perspective transmission coding step According to the obtained series of point cloud segments that are divided into corresponding perspective categories, the perspective transmission scheme is adopted, and high-definition coding transmission is performed for the perspective that needs to be observed, and for perspectives that do not need to be observed Perform lower-definition coding transmission.
- Point cloud segment determination step segmentation is performed according to the normal vector of the point cloud, and the distance between the normal vector of each point in the point cloud and the 6 standard normal vectors is calculated, and the standard normal vector with the closest distance is taken as the point category.
- the 6 standard normal vectors adopt the existing technology, which are: (1, 0, 0), (-1, 0, 0), (0, 1, 0), (0, -1, 0), (0, 0, 1), and (0, 0, -1).
- the extraction of the connected components includes: the point cloud fragments of the six major categories, according to the spatial distance between points and the adjacency relationship between points , Divide the point cloud fragments where the spatial distance between the points is greater than the preset distance range and the points are not adjacent to each other into smaller point cloud fragments, so that the point cloud fragments after the re-division are between the points
- the spatial distance of is smaller than the preset distance range and the points are adjacent to each other;
- the re-divided point cloud fragments still belong to 6 categories corresponding to the 6 standard normal vectors respectively, and then the obtained re-divided point cloud fragments are projected in different directions according to the normal vector categories;
- Projection transmission schemes that perform sub-view transmission coding include any one or more of the following: a scheme that does not modify the point cloud projection method, a scheme that improves the point cloud projection method, and a standard normal vector scheme that is modified according to the user's perspective.
- FIG. 1 is a process diagram of processing point cloud fragments from perspectives without modifying the point cloud projection solution in the first embodiment. It can be seen from Fig. 1 that the 3D point cloud object is divided according to 6 standard vectors to obtain point cloud segment 1 to point cloud segment t.
- the point cloud fragments that need to be mainly observed and the point cloud fragments that do not need to be mainly observed are divided according to the observation angle of view.
- the patch index list confirm which 2D segments are projected to be mainly observed, and which do not need to be mainly observed.
- the selection of the standard vector projection plane as 0, 1, and 2 is the one that needs to be mainly observed.
- the corresponding point cloud segments 1, 3, 6, 7, 9, etc. are divided into the point cloud segments that need to be mainly observed, and the rest are divided into the point cloud segments that do not need to be mainly observed Point cloud fragment.
- Figure 3 is an example diagram of different faces that need to be mainly observed under different user perspectives in the first embodiment; for a point cloud object, among the six faces of its three-dimensional bounding box, Figure 3 shows one There are three situations in which the surface corresponding to the normal vector needs to be mainly observed, the surface corresponding to the two normal vectors needs to be mainly observed, and the surface corresponding to the three normal vectors needs to be mainly observed.
- the point cloud projection transmission method based on the user’s perspective of the present invention includes any one of the following:
- the 1, 2, and 3 improved point cloud projection schemes include:
- Observation steps of main observation fragments According to the obtained series of point cloud fragments that are classified into corresponding perspective categories, according to the observation perspective, they are divided into the point cloud fragments that need to be mainly observed, and the point cloud fragments that do not need to be mainly observed, which is confirmed Which standard normal vector corresponds to the point cloud segment that needs to be mainly observed.
- Figure 2-1 is a process diagram of processing point cloud fragments from the first perspective under the improved point cloud projection scheme in the first embodiment.
- the first improved point cloud projection scheme is the same segment division process as in Figure 1.
- Figure 2-1 of this embodiment after confirming which standard normal vector corresponds to the point cloud segment that needs to be mainly observed, The point cloud fragments that need to be mainly observed and the point cloud fragments that do not need to be mainly observed are projected onto different planes as needed. For example, the point cloud fragments that need to be mainly observed are projected into picture A, and they do not need to be projected.
- the main observation point cloud fragment is projected into picture B, and the size of picture A and picture B are changed, and after the enlargement or reduction process is performed, in this embodiment, picture C is obtained after picture B is reduced, and picture A remains unchanged .
- the unobserved picture C is merged into the observable picture A to obtain a projected picture, which is then compressed and encoded.
- the present invention it is within the scope of the present invention to change the size of the picture, such as zooming in picture A while picture B remains the same or zoomed out, or picture A remains unchanged while reducing picture B, which is within the scope of the invention, making it unnecessary It is sufficient if there is a significant size difference between the observation picture and the picture that needs to be observed, and the combination scheme of size change and the magnification and reduction coefficients do not limit the scope of the invention.
- the actual shape of the generated point cloud segment is based on the actual situation. In the figure, only the block diagram represents the point cloud segment for illustration.
- Fig. 2-2 is a process diagram of processing point cloud fragments from a second perspective under the improved point cloud projection scheme in the first embodiment
- the second improved point cloud projection scheme After confirming which standard normal vectors correspond to the point cloud fragments that need to be mainly observed, it is the same as the first improved point cloud projection scheme above, and the point cloud fragments are projected to On different planes, the point cloud fragments that need to be mainly observed are projected into picture A, and the point cloud fragments that do not need to be mainly observed are projected into picture B. In this case, no stitching is performed, but the pictures that need to be observed are based on requirements. A. There is no need to observe the pictures to be resized separately, for example, after the size is enlarged or reduced, the projected pictures after the resize are respectively subjected to subsequent compression coding.
- Figure 2-3 is a process diagram of processing point cloud fragments from the third perspective under the improved point cloud projection scheme in the first embodiment
- the third improved point cloud projection scheme is to project point cloud segment 1 to point cloud segment t on six different planes according to their standard normal vector types, corresponding to the six standard vectors, and get the first group of point clouds. Fragments to the sixth group of point cloud fragments. After confirming which standard normal vector corresponds to the point cloud fragments that need to be observed, the size of the planes corresponding to these point cloud fragments is changed according to requirements, as shown in Figure 2-3. According to the user’s perspective, select the point cloud segment on the projection surface of the standard vectors 0, 1, 2 as the main observation surface, that is, the point cloud segments of the 1-3 groups remain unchanged, and the point cloud segments of the 4-6 groups are reduced. , And then compress and code separately after such processing.
- Observation angle normal vector calculation steps Obtain the direction of the user's observation angle, and calculate the normal vector of the observation angle. The specific process is as follows:
- the rest of the standard normal vector calculation steps define the normal vector as one of the standard normal vectors, that is, the normal vector and one of the standard normal vectors are rotated and mapped, and then the other standard normal vectors are subjected to the same rotation mapping processing to make the overall Under the premise that the modification amount of is small, to obtain the other corresponding five sets of standard normal vectors, the calculation process is as follows:
- FIG. 4 is a schematic block diagram of the system of the sending end in the first embodiment
- the present invention provides a point cloud projection transmission system based on a user's perspective, which includes a segment generation module, a point cloud segment determination module, and a perspective transmission encoding module as a projection encoding processing module.
- the fragment generation module is used to preliminarily divide the point cloud object on the standard vector to obtain the point cloud fragment to be processed.
- Point cloud segment determination module Determine the corresponding point cloud segment according to the observation perspective, and obtain a series of point cloud segments that are classified into the corresponding perspective category;
- the perspective transmission coding module is used to determine the projection transmission scheme according to the user perspective, and process the to-be-processed point cloud segment. According to the obtained series of point cloud fragments divided into corresponding viewing angle categories, the perspective transmission scheme is adopted to perform high-definition encoding transmission for the viewing angle that needs to be observed, and lower-definition viewing angle for the viewing angle that does not need to be observed. Encoding transmission.
- the point cloud segment determination module :
- the 6 standard normal vectors are:
- connection component extraction refers to: taking 6 broad categories of point cloud fragments, according to the spatial distance between the points and the adjacency relationship between the points, the spatial distance between the points is greater than the preset distance range And the point cloud fragments that are not adjacent to each other are divided into smaller point cloud fragments, so that the spatial distance between the points of the re-divided point cloud fragment is less than the preset distance range and the points are adjacent to each other ;
- the re-divided point cloud fragments still belong to 6 categories corresponding to the 6 standard normal vectors respectively, and then the obtained re-divided point cloud fragments are projected in different directions according to the normal vector categories;
- the sub-view transmission coding module :
- the perspective transmission solution includes any one or more of the following: a solution that does not modify the point cloud projection method, a solution that improves the point cloud projection method, and a standard normal vector solution that is modified according to the user's perspective.
- a solution that does not modify the point cloud projection method a solution that improves the point cloud projection method
- a standard normal vector solution that is modified according to the user's perspective.
- the present invention provides a point cloud receiving and processing system based on the user's perspective, including:
- the inverse processing module uses the codec parameters determined from the user's perspective to obtain a video sequence from the received bitstream, process the video sequence to obtain a point cloud segment, and decode and reconstruct each point cloud segment to obtain a point cloud object.
- the point cloud projection transmission system based on the user's perspective and the point cloud receiving and processing system based on the user's perspective provided by the present invention can pass the point cloud projection transmission system method based on the user's perspective and the point cloud receiving processing method based on the user's perspective provided by the present invention The steps of the process are realized.
- the purpose of the present invention is to provide a point cloud projection transmission scheme based on the user's perspective, which is modified and perfected on the basis of the existing typical point cloud compression scheme, and provides different degrees of observation at different viewing angles according to the needs of users.
- the rendering effect is to provide a point cloud projection transmission scheme based on the user's perspective, which is modified and perfected on the basis of the existing typical point cloud compression scheme, and provides different degrees of observation at different viewing angles according to the needs of users.
- the rendering effect is to provide a point cloud projection transmission scheme based on the user's perspective, which is modified and perfected on the basis of the existing typical point cloud compression scheme, and provides different degrees of observation at different viewing angles according to the needs of users. The rendering effect.
- the present invention is implemented based on the current classic point cloud sequence coding transmission scheme.
- the classic point cloud sequence coding method needs to be carried out sequentially. Segments are divided according to normal vectors, segment is subdivided according to segment connections, and projections in different directions are performed according to normal vectors. Fill the projection plane with blanks, and use a two-dimensional coding tool such as HEVC to compress and encode the video sequence obtained by the projection.
- the data corresponding to the viewing angle that needs to be observed can be preferentially allocated to ensure that the data corresponding to the viewing angle that needs to be observed has sufficient bandwidth for transmission.
- the data corresponding to the viewing angle that needs to be observed can be allocated according to a higher ratio, and the data corresponding to the viewing angle that does not need to be observed according to a lower ratio.
- the first solution is based on the classic point cloud sequence coding, after obtaining the video sequence after projection, without changing the structure of the video sequence, it is achieved by modifying the coding parameters of the video coding tool. Specifically, because the point cloud sequence is divided into different fragments according to the normal vector, we can find the normal vector corresponding to the corresponding perspective according to the perspective, and further find the corresponding fragments. For these fragments, locate in the two-dimensional picture The corresponding projected segments can then be assigned different compression parameters to these segments in the video encoding tool. Different compression parameters can achieve different compression coding effects. For example, for the compression parameter quantization step sequence number, QP value for short, a low QP value can achieve higher compression quality, and a high QP value can achieve lower compression quality. .
- the projection mode needs to be modified, so the structure of the video sequence will be modified accordingly, but the two-dimensional video sequence is still obtained and compressed using a two-dimensional coding tool.
- a two-dimensional coding tool For example, after segment division and projection are performed on the point cloud, down-sampling processing can be performed on the two-dimensional segment obtained from the point cloud segment of the unwanted perspective. After such processing, we can obtain a video sequence with composite resolution, and then use a two-dimensional coding tool to compress with the same resolution and accuracy.
- the standard normal vector can be modified according to the user's perspective.
- the current classic compression algorithm of dynamic point cloud is to divide the divided point cloud fragments into categories according to the relationship between the six standard normal vectors, and then select one of them as the main projection direction and project it onto the plane. Based on this projection method, the standard normal vector can be modified according to the user's perspective, so that the user's perspective corresponds to the front of the projection.
- the three schemes are not opposed to each other and can be used in combination.
- the standard normal vector can be modified according to the user's perspective, and then the projection mode or coding parameters can be modified.
- the present invention adopts three solutions.
- the three schemes implement split-view coding in terms of different strategies, and they can all be combined with each other, or the three can be used in combination.
- the corresponding point cloud segment needs to be determined according to the observation angle of view. It can be implemented relatively simply on the basis of the current point cloud sequence coding method. After the step of determining the point cloud segment corresponding to the perspective, a series of point cloud segments classified into the corresponding perspective category can be obtained. This is the premise of the three technical solutions in the point cloud sub-perspective transmission strategy proposed in this patent. And foundation.
- the specific implementation algorithm is as follows.
- the current classic coding method of point cloud sequence is to divide the segment according to the normal vector of the point cloud. By calculating the distance between the normal vector of each point in the point cloud and the 6 standard normal vectors:
- connection component extraction processing that is, points that have been divided into six categories, according to the spatial distance and the adjacency relationship between points, are divided into smaller point cloud segments to prevent points from appearing in the same segment. The distance is too large and not adjacent.
- a parameter index can be assigned to indicate which normal vector plane they belong to.
- Table 2 is the patch index list in the first embodiment. ):
- the method of confirming the corresponding normal vector from different perspectives can include the following:
- the standard normal vector (0, 1, 0) corresponds to the perspective from the front.
- the normal vector belongs to (0, 1). , 0) Those point cloud fragments of this category. If it is not observed from a standard forward perspective, there are more types of point cloud fragments that are mainly observed.
- the two-dimensional pictures that need to be compressed are divided into blocks, that is, the two-dimensional segments are divided evenly, and then the QP offset values and other codes are set for these blocks respectively.
- the parameters are encoded.
- a parameter index can be assigned to each point cloud segment, and then a parameter list patch index list can be assigned to each block of the picture that needs to be compressed and encoded to record which point cloud segments this block is associated with.
- these point cloud segments have been divided into point cloud segments that need to be primarily observed and point cloud segments that do not need to be primarily observed according to the observation angle of view. Therefore, according to the relationship between the point cloud segment and the two-dimensional segment, that is, the patch index list, it is possible to confirm which of the projected two-dimensional segments need to be mainly observed and which do not need to be mainly observed.
- each segment is encoded and decoded.
- the point cloud projection method can also be modified to achieve the coding effect of the point cloud by viewing angle and resolution.
- the method for projecting each point cloud segment to a two-dimensional picture is as follows: Projection is performed according to the standard normal vector category of each point cloud segment, for example, the standard method of a point cloud segment The vector category is (1, 0, 0), then when projecting, the (y, z) coordinates in the three-dimensional coordinates (x, y, z) of this point cloud segment are used as plane coordinates, and the (x) coordinates are used Make the projection depth. As for where the projection position is on the picture, this is mainly based on the way to save the overall occupied space, starting from the upper left corner, filling into the two-dimensional picture in turn, so that the final total occupied plane size is the most economical.
- the present invention proposes to optimize the way of projection of the point cloud from different perspectives. And based on this idea, the following three optimization ideas are proposed.
- the three optimization ideas can be performed in parallel, that is, you can choose one:
- the point cloud fragments are projected onto different planes as needed, processed, and then spliced onto a picture.
- the point cloud fragments that need to be mainly observed are projected onto picture A according to the classic method, and the size is 1280 ⁇ 1280
- the point cloud fragments that do not need to be mainly observed are projected onto picture B according to the classic method, and the size is 1280 ⁇ 1280 .
- the picture B is reduced to obtain the picture C, and the size is 640 ⁇ 640. Since the projection according to the classic method ensures the minimum total occupied space, the projected position on the picture A is likely to be no more than 1/4, so the picture C can be pasted in the area of the picture A that is not occupied by the projection.
- the picture occupies a size of 1280 ⁇ 1280 before the optimization, and the picture occupies a size of 1280 ⁇ 1280 after the optimization, but the plane space occupied by the point cloud fragments not mainly observed is smaller.
- the coding technology can achieve the purpose of dividing the definition coding presentation according to the viewing angle requirements.
- the point cloud segments are projected onto different planes as required, and these planes are processed according to requirements before compression and encoding.
- point cloud fragments that need to be mainly observed are projected onto picture A according to the classic method
- point cloud fragments that do not need to be mainly observed are projected onto picture B according to the classic method
- picture B is reduced in size. After that, both the picture A and the reduced picture B are compressed and encoded for presentation.
- the video sequence presented before optimization is a set of video sequences
- the video sequence presented after optimization is two groups of video sequences.
- the point cloud fragments that are not mainly observed are reduced due to the reduction process, which reduces the difficulty and cost of compression encoding, and adopts the same video
- the coding technology can achieve the purpose of dividing the definition coding presentation according to the viewing angle requirements.
- each point cloud segment is projected onto planes A, B, C, D, E, and F according to their standard normal vector categories, respectively, according to the classic method. Since it is possible to determine which standard normal vector point cloud fragments need to be mainly observed according to the user's perspective, it is possible to determine which of the six planes do not need to be mainly observed, and can be reduced in size. After performing this processing on each frame of the point cloud sequence, six groups of video sequences can be obtained, and the same video coding technology is used for coding processing.
- the video sequence presented before optimization is a group of video sequences
- the video sequence presented after optimization is multiple groups of video sequences.
- the point cloud fragments that are not mainly observed are reduced due to the reduction process, which reduces the difficulty and cost of compression encoding.
- Video coding technology can achieve the purpose of dividing definition coding and presentation according to viewing angle requirements.
- the second method the method of modifying the point cloud projection method is realized by the following.
- Solution 1 A solution in which segments corresponding to the user's perspective and segments corresponding to the non-user's perspective are spliced on a video sequence. It is necessary to record the index of all segments corresponding to the non-user perspective and the size reduction ratio at the sending end. This information will be encoded and transmitted as new additional information; at the receiving end, it needs to be based on the received index information and the reduction ratio. Restore these fragments to the original size, and then perform the subsequent 3D point cloud reconstruction steps.
- Solution 2 A solution of projecting segments corresponding to the user's perspective and segments corresponding to the non-user's perspective to two video sequences respectively.
- each video sequence needs to be restored according to the position of each video sequence in the code stream, and then these video sequences need to be restored to the original size according to the received reduction ratio, and then the subsequent steps Three-dimensional point cloud reconstruction steps, and need to stitch together the point clouds recovered from the six groups of video sequences.
- the standard normal vectors for dividing point cloud segments are the following six normal vectors:
- the main observing projection direction may be too much.
- the main projected surface to be observed needs to include the front, side, and top of the observed object, that is, there are at least three sides.
- the present invention proposes to adaptively adjust the direction of the standard normal vector according to the user's observation angle of view.
- the direction of the user's observation angle can be obtained, and the normal vector of the observation angle can be calculated.
- the normal vector of the observation angle can be calculated. For example, when observing a point cloud object, take the observed point cloud object as the origin of the relative coordinate system, provide the user with the relative position coordinates of the observed point cloud object, and calculate the position difference between the two to obtain the relative position.
- the normal vector of the direction of the viewing angle which represents the direction of the user's viewing angle.
- the normal vector is defined as one of the standard normal vectors, that is, the normal vector and one of the standard normal vectors are rotated and mapped, and then the other standard normal vectors are subjected to the same rotation mapping processing, so that the overall modification amount is small Under the premise, to obtain the other corresponding five sets of standard normal vectors.
- the calculation formula i is as follows:
- the normal vector has been normalized, that is, the sum of the squares of the three coordinates is 1.
- the other five standard normal vectors (0, 1, 0), (-1, 0, 0), (0, -1 , 0), (0, 0, 1), (0, 0, -1) corresponding to the normal vector general formula of the rotation map can be expressed as:
- the method to ensure that the overall amount of modification is small is to make the new six standard normal vectors, including the user’s perspective normal vector, correspond to the corresponding six standard methods after the user’s perspective normal vector corresponds to the standard normal vector and is rotated.
- the sum of the dot products of the vectors is as large as possible.
- the general formula of the normal vector of the user's viewing angle Set it to the standard normal vector (1, 0, 0) as a rotation mapping relationship
- the other five standard normal vectors (0, 1, 0), (-1, 0, 0), (0, -1 , 0), (0, 0, 1), (0, 0, -1) corresponding to the normal vector general formula of the rotation map can be obtained by calculation:
- segment projection and video coding can adopt the other two modification schemes mentioned in the present invention.
- method three the method of modifying the standard normal vector.
- the new standard normal vector needs to be recorded at the sending end, and this part of the information will be encoded and transmitted as additional additional information; at the receiving end, the inverse process of point cloud projection needs to be performed according to the new standard normal vector, so as to start from the two-dimensional video
- the sequence reconstructs a three-dimensional point cloud sequence.
- the transmission strategy of the present invention adaptively adjusted according to the user's perspective, there is no need to make major modifications to the current existing point cloud transmission strategy, and the user's attention to various aspects is also different, distinguishing between the point cloud fragments that need to be observed and the point cloud fragments that do not need to be observed.
- the observed point cloud fragments can be better combined with the current high compression rate encoding algorithm, without losing the experience quality of the user's main perspective, and without affecting the compression rate and the objective quality of the point cloud, saving transmission consumption The resources to achieve the purpose of sub-view transmission.
- the processing steps of the point cloud sequence encoding method include: dividing the point cloud object into multiple point cloud segments according to the normal vector, subdividing the segments according to the connection relationship of the point cloud segments, and dividing the subdivided point cloud
- the segments are projected according to the normal vector, the projection plane is blank filled, and the video sequence obtained by the projection is compressed and encoded using a two-dimensional coding tool such as HEVC.
- the present invention is implemented by improving on the existing point cloud sequence coding transmission scheme, and the user’s viewing angle is introduced in the process of compressing the three-dimensional point cloud into a two-dimensional image to optimize and adjust the point cloud projection processing process. Thereby improving compression performance.
- the projection vector of the point cloud segment is selected by comprehensively considering the two factors of the user’s viewing angle and the normal vector of the segment.
- This embodiment provides a point cloud projection processing method based on the user’s perspective, including: distinguishing points based on the user’s perspective The main segment and multiple edge segments of the cloud object; considering the spatial connection between the edge segment and the main segment, the projection vector of the edge segment is maintained or changed based on the user's perspective; subsequent point cloud processing is performed after projection.
- FIG. 5 is a three-dimensional schematic diagram of a three-dimensional edge box of a point cloud object in the second embodiment
- the point cloud object is projected by using a three-dimensional bounding box of the point cloud object to include the point cloud object.
- a three-dimensional bounding box of a point cloud object is a rectangular cube set for a specific point cloud object, and the specific size and shape are not limited by this embodiment.
- a vector pointing from the point cloud to the user's observation point is used to represent the user's view of the user's observation, which is called the view vector. Then calculate the distance between the viewing angle vector and the six standard normal vectors, and the standard normal vector with the closest distance to the viewing angle vector is called the principal vector under the viewing angle.
- Table 3 is a patch index list of index parameter values and projection vector values in the second embodiment, and their parameter assignments and corresponding
- the standard projection vector surface is the same as Table 2.
- 0 is used to refer to the direction of the first standard normal vector, which is perpendicular to the (1, 0, 0) plane
- 1 is used to refer to the direction of the sixth standard normal vector, which is perpendicular to the (-1, 0, 0) plane
- use 2 to refer to the direction of the second standard normal vector, perpendicular to the (0, 1, 0) plane
- use 3 to refer to the direction of the third standard normal vector, perpendicular to the (0, -1, 0) ) Plane
- 4 to refer to the direction of the fourth standard normal vector, perpendicular to the (0, 0, 1) plane
- 5 refer to the direction of the fifth standard normal vector, perpendicular to the (0, 0, -1) surface.
- the main vector is the first standard normal vector (1, 0, 0).
- the six standard normal vectors have a one-to-one correspondence with the six faces of the three-dimensional edge box, and the standard normal vectors are perpendicular to their corresponding faces.
- Index parameter value Projection vector surface Normal 0 (1, 0, 0) 1 (-1, 0, 0) 2 (0, 1, 0) 3 (1, -1, 0) 4 (0, 0, 1) 5 (0, 0, -1)
- the closest standard normal vector is the The standard normal vector corresponding to the point.
- the points in all the point cloud objects are divided into 6 first-level point cloud segments.
- the 6 first-level point cloud segments are subdivided, and the points that belong to the same first-level point cloud segment and are relatively close are divided into a second-level point cloud segment.
- the method of calculating the distance between points can be selected according to the specific situation. Euclidean distance is usually used. According to this method, the 6 first-level point cloud segments are subdivided, and the normal vectors of the second-level point cloud segments obtained after subdivision all correspond to one of the above-mentioned 6 standard normal vectors.
- the secondary point cloud segment is used as the cloud segment to be processed to continue processing, and the cloud segments to be processed on the six standard normal vector planes are further distinguished according to the direction of the view angle vector. Calculate the normal vector of the secondary point cloud segment. Then the main direction of the projection is determined according to the user's viewing angle. According to the relationship between the normal vector of the secondary point cloud segment and the main direction, the secondary point cloud segment is divided into the main segment, the edge segment, and the irrelevant segment.
- FIG. 6 is a diagram of the projection adjustment process of the point cloud segment under the projection optimization adjustment scheme in the second embodiment
- point cloud segment 1 to point cloud segment t are projected onto six different planes according to their standard normal vector types, corresponding to the six standard vectors, and the first 1 set of point cloud fragments to the 6th group of point cloud fragments, and then determine the main direction of projection according to the user’s viewing angle.
- the secondary point cloud fragments are these 6 groups
- Point cloud fragments are divided into main fragments, edge fragments, and irrelevant fragments.
- the first group is divided into main segments
- the second group is divided into irrelevant segments
- the 3-6 groups are divided into edge segments.
- projection is performed according to the optimized and adjusted projection direction.
- the main segment and the irrelevant segment keep the projection direction unchanged, adjust the edge segment according to the large or small spatial correlation, and obtain the first group of point cloud segments to the sixth group of point clouds after the projection direction is optimized and adjusted. Fragments, it is obvious that the first group of point cloud fragments towards the user's perspective are more concentrated.
- Fig. 7 is a corresponding table of the relationship between normal vectors and principal vectors of different point cloud segment types in the embodiment of the present invention; Fig. 7 shows the distinguishing criteria for different segment types.
- the secondary point cloud fragments with the same normal vector and the main vector are defined as the main fragments. These main fragments are also called the first standard fragments. These main fragments form the main fragment set, the first standard fragment The set (ie, the main fragment set) is gathered on the main fragment projection gathering surface of the point cloud 3D edge box.
- the secondary point cloud fragments whose normal vector is opposite to the main vector as irrelevant fragments. These irrelevant fragments are also called second standard fragments. These irrelevant fragments form a set of irrelevant fragments, and the second standard fragment set (ie, irrelevant point cloud fragments) Set) is gathered on the projection gathering surface of the irrelevant fragments of the 3D edge box of the point cloud.
- edge fragments The secondary point cloud fragments whose normal vector is perpendicular to the main vector are defined as edge fragments. These edge fragments are distributed on the projection gathering surface of the four edge fragments perpendicular to the main vector.
- Figure 5 shows only four edge fragments. One of the projection surfaces is projected, and the other three surfaces are not marked.
- these secondary point cloud segments located on the projection gathering surface of the first edge segment that is, these edge segments form a set of the first edge segment, and the first edge segment is derived from the third standard segment.
- the second-level point cloud segments located on the gathering surface of the other three edge segments are called the second edge segment set, the third edge segment set, and the fourth edge segment set. The above forms a collection, from the 4th to 6th standard fragments.
- the projection vector of the main segment is projected with its standard normal vector, that is, the main vector.
- the projection vector of the irrelevant segment is projected with its standard normal vector, that is, the opposite of the principal vector.
- the size of the spatial connection is judged by sequentially calculating the intersection ratio (IOU) of each edge segment with each main segment. If there is an IOU greater than the threshold, it means that the edge point cloud segment has a greater spatial connection with the main point cloud segment, and the main vector is selected as the projection vector of the edge point cloud segment; if there is no IOU greater than the threshold, it means the edge point The cloud segment has a small spatial connection with the main point cloud segment, and the normal vector of the edge point cloud segment is selected as its projection vector.
- IOU intersection ratio
- Figure 8-1 is a schematic diagram of the relative positions of the three-dimensional edge box of the main segment and the three-dimensional edge box of the edge segment used for projection adjustment according to the main vector direction in the second embodiment
- Figure 8-2 is the main segment and the three-dimensional edge box in Figure 8-1
- a schematic diagram of the two-dimensional projection obtained after the three-dimensional edge box of the edge segment is projected onto the two-dimensional plane; the IOU judgment process of one of the edge segments and a certain main segment of each main segment is described in detail.
- Fig. 8-1 is a schematic diagram of the relative positions of the three-dimensional edge box of the main segment and the three-dimensional edge box of the edge segment for projection adjustment according to the main vector direction in the embodiment of the present invention.
- the three-dimensional edge box (called edge box 1) shown by the solid line corresponds to the main segment
- the three-dimensional edge box (called edge box 2) shown by the dashed line corresponds to the edge segment.
- Figure 8-1 and Figure 8-2 are only a brief enumeration of the three-dimensional edge boxes of a certain main segment and a certain edge segment, showing the projection orientation between the three-dimensional boxes and the overlap of the two-dimensional projection map.
- Figure 8-1 and Figure 8-2 are only a brief enumeration of the three-dimensional edge boxes of a certain main segment and a certain edge segment, showing the projection orientation between the three-dimensional boxes and the overlap of the two-dimensional projection map.
- one-by-one overlapping calculation spatial associations between a certain edge segment and other main segments, or between other edge segments and other main segments it will not be listed one by one, but only an example.
- Fig. 8-2 is a rectangular two-dimensional projection diagram obtained after the three-dimensional edge boxes of the main segment and the edge segment are projected onto a two-dimensional plane.
- the three-dimensional edge box of the main segment is projected along the main vector direction to obtain the main two-dimensional projection image
- the three-dimensional edge box of the edge segment is projected along the main vector direction to obtain the edge two-dimensional projection image.
- the pattern filling part is projected along the principal vector direction of the edge box 1 of the main segment (called rectangular area 1)
- the dot filling part is projected along the principal vector direction of the edge box 2 of the edge segment (called rectangular).
- Area 2), the grid point filling part is the common part where the rectangular area 1 and the rectangular area 2 overlap (referred to as the rectangular area 3).
- S 3 corresponds to the area of the rectangular area 3 in Figure 4, that is, the overlapping common part of the main two-dimensional projection image and the edge two-dimensional projection image
- S 1 corresponds to the main two-dimensional projection map in Figure 4 excluding the area of the common part
- S 2 corresponds to the two-dimensional projection of the edge in Fig. 4 excluding the area of the common part.
- a i and B i are the three-dimensional edge boxes of the main segment, that is, the vertices closest to the origin O on the edge frame 1; as shown in Figure 8-1
- a j and B j are respectively the three-dimensional edge box of the edge segment, that is, the vertices of the edge box 2 closest to the origin O and the farthest.
- a 'i, A' j, B 'i, B' j are A i, A j, B i , B j of the resultant vector along the main direction of the projection point.
- Their coordinates in the XYZ coordinate system with O as the origin are: A i (x i0 ,y i0 ,z i0 ),B i (x i1 ,y i1 ,z i1 ),A j (x j0 ,y j0 ,z j0 ),B j (x j1 ,y j1 ,z j1 ),A′ i (y i0 ,z i0 ),B′ i (y i1 ,z i1 ),A′ j (y j0 ,z j0 ) ,B′ j (y j1 ,z j1 ).
- For an edge segment compare its respective IOUs with an intersection ratio threshold. If there is an IOU greater than the intersection ratio threshold, it means that the edge segment has a greater spatial connection with the main segment, in order to optimize the point of the main vector direction Cloud visual quality, change the projection direction of the edge segment, and select the main vector as the projection vector of the edge segment; if there is no IOU greater than the intersection ratio threshold, it means that the edge segment has a small spatial connection with the main segment, then keep this The normal vector of the edge segment is used as its projection vector.
- the final projection vector of the edge segment is determined by the normal vector of the edge segment and the spatial position relationship between the edge segment and the main segment.
- the selection of the intersection ratio threshold depends on the surface characteristics of the point cloud object to be processed. For example, if the surface characteristics of the object point cloud object and the character point cloud object are different, then the intersection ratio threshold value selected when comparing It's different.
- the threshold value is generally selected to be 0.2-0.3, and further optionally, the threshold value is determined mainly through adjustment and selection of experimental results.
- the adjustment, optimization or maintenance of the projection vector is completed for the first edge segment, and the same processing of the above process is used to discriminate and adjust each edge segment one by one, until the second edge segment is set and the third edge segment is set. Set, all edge fragments in the set of 4th edge fragments have been processed.
- the projection relationship identification field is used to identify: the projection vector of the point in the point cloud object after processing and adjustment, including the optimized and adjusted projection vector coordinate value and the corresponding index parameter value.
- Table 4 is an example table of the corresponding relationship before and after adjustment and optimization of the projection vectors of multiple edge segments.
- the main segment is: the index parameter value is 0, the standard normal vector coordinate value is (1, 0, 0) without optimization adjustment processing, and the main vector is maintained for projection.
- Irrelevant edge fragments are: index parameter value 1, standard normal vector coordinate value (-1, 0, 0) without optimization and adjustment processing, keeping the original standard normal vector direction for projection.
- 2nd edge segment 2 (0, 1, 0) Not big 2 (0, 1, 0) 3rd edge segment 4 (0, 0, 1) Big 0 (1, 0, 0) 4th edge segment 5 (0, 0, -1) Not big 5 (0, 0, -1) 5th edge segment 3 (0, -1, 0) Big 0 (1, 0, 0) K-th edge segment ... ... ... ... ... ...
- the edge segment that has undergone projection processing, and the irrelevant segment project them according to the adjusted projection vector to generate a compressible two-dimensional image, and further process to obtain a two-dimensional video sequence, and use a video codec for the two-dimensional video sequence.
- the video sequence is compressed.
- FIG. 9 is a schematic flowchart of a point cloud projection transmission method based on a user's perspective in an embodiment of the present invention. the steps of a point cloud projection processing method based on a user's perspective include:
- the point cloud object is divided to generate first-level point cloud fragments.
- the first-level point cloud fragments are preliminarily divided according to the difference of the six standard vectors;
- the process of generating a secondary point cloud segment from the primary point cloud segment can be completed by using the usual technical means in the existing point cloud projection step, and the secondary point cloud segment is used as a cloud segment to be processed and the following steps are performed.
- obtaining cloud fragments to be processed is not limited to the above-mentioned steps such as the first-level point cloud fragment and the second-level point cloud fragment.
- the preprocessing steps of any point cloud fragment in the prior art are all within the scope of the present invention. .
- the cloud fragments to be processed are distinguished. According to the relationship between the normal vector of the secondary point cloud fragment and the main vector, the secondary point cloud fragment is divided into the main fragment, the edge fragment, and the irrelevant fragment. Among them, the normal vector and the main vector are the same. Fragment, normal vector and main vector opposite are irrelevant fragments, normal vector and main vector are perpendicular to edge fragments;
- the projection direction of the secondary point cloud segment is calculated, and the projection direction of the edge segment in the secondary point cloud segment is optimized and adjusted.
- the specific optimization and adjustment process is as described above;
- FIG. 10 is a schematic flowchart of a point cloud receiving and processing method based on a user's perspective in an embodiment of the present invention.
- the point cloud receiving and processing method based on the user's perspective in this embodiment is the inverse process of the above-mentioned point cloud projection processing method based on the user's perspective, and the decompression on the receiving end is the inverse process of compression on the transmitting end.
- the first-level point cloud fragment is combined with auxiliary information to restore and generate a point cloud object, and the above-mentioned projection relationship identification field is used for back projection processing to restore the two-dimensional point cloud fragment to a three-dimensional point cloud object;
- FIG. 11 is a schematic block diagram of the point cloud projection processing system based on the user's perspective in an embodiment of the present invention.
- the point cloud projection processing system based on the user's perspective includes a point cloud input module, a segment generation module, a segment division module, a projection vector adjustment module, a projection module and a compression module, and a code stream output module.
- Point cloud input module used to input point cloud objects
- the fragment generation module is used to generate cloud fragments to be processed from the point cloud object according to predetermined rules, and perform subsequent projection optimization adjustments;
- Fragment division module used to divide the cloud fragment to be processed into main fragments, edge fragments and irrelevant fragments
- the projection vector adjustment module is used to take into account the spatial connection between the edge segment and the main segment, and to maintain the projection vector of the edge segment according to the above projection vector adjustment rule or change it based on the user's perspective;
- Projection module used for projecting the main segment, edge segment and irrelevant segment after projection adjustment
- the compression module and the code stream output module are used to compress the two-dimensional image after projection; and the subsequent code stream output.
- the projection module uses the optimized and adjusted projection vector for projection, and is identified by the projection relationship identification field.
- the projection relationship identification field contains the index parameter value of the optimized and adjusted projection vector and the corresponding projection vector value .
- the point cloud receiving and processing system based on the user's perspective includes: a code stream processing module and a back projection module, a code stream processing module, used to receive the input of the code stream; a back projection module, used to back project two-dimensional point cloud fragments, according to The indication of the projection relationship identification field restores the two-dimensional point cloud segment to a point cloud object.
- the projection relationship identification field contains the index parameter value of the optimized and adjusted projection vector and the corresponding projection vector value.
- each module in the point cloud projection processing system based on the user's perspective and the point cloud receiving and processing system based on the user's perspective are the same as the above-mentioned point cloud projection processing method based on the user's perspective and the point cloud receiving based on the user's perspective.
- the optimization and adjustment process in the processing method corresponds to the corresponding, which can be deduced by analogy. Then the structure and technical elements of the device can be formed by the corresponding conversion of the generation method and the receiving method. The repeated description is omitted here, and the details are not repeated here.
- the point cloud segment corresponding to the same standard vector may have a large spatial distance between the points contained.
- Such a simple rough division operation has nothing to do with the user's perspective, and the point cloud segment corresponding to the same standard vector is geometrically spatially related Small and fragmented.
- the associated points in the point cloud object that are very close in space are segmented and projected into different standard vectors because of the difference in standard vectors.
- the factor of the user's perspective is introduced to perform the multiple point cloud fragments.
- the user has a high degree of attention to the front content of the point cloud object, then the main segment is determined based on the user's perspective.
- the edge segments associated with the main segment such as the left and right sides and the top and bottom sides Point cloud segments are used as edge segments.
- the projection direction of these edge segments further considers the distance relationship between these edge segments and the main segment.
- the projection vector of the points in the edge segment is adjusted and optimized.
- the projection vector of the point in the edge segment is changed based on the user's perspective, so that it is projected according to the main vector direction consistent with the main segment.
- the spatial connection is small, keep the edge segment The projection vector of the point.
- the above optimization and adjustment of the projection direction of the edge segment takes into account the user's perspective and the spatial correlation with the main vector, so that those points in the edge segment that are largely related to the main segment are changed and projected according to the main vector of the main segment, which pays more attention to the user.
- the direction of higher degree of concentration is concentrated, and the adjusted and concentrated points are originally highly related and close in space with the main segment of higher attention. This avoids the scattered fragmentation and overcomes the problem of point cloud objects.
- the associated points with very close spatial distance are divided and projected into different standard vectors because of the difference in standard vectors.
- the present invention also provides a point cloud projection transmission method based on the user's perspective, a point cloud receiving processing method based on the user's perspective, and a point cloud projection transmission system based on the user's perspective, and a point cloud projection transmission system based on the user's perspective.
- Point cloud receiving and processing system a point cloud projection transmission method based on the user's perspective, a point cloud receiving processing method based on the user's perspective, and a point cloud projection transmission system based on the user's perspective.
- the above-mentioned first embodiment and the second embodiment are combined as the technical solution of the third embodiment.
- There are three types in this embodiment and the specific content of the projection adjustment optimization and the multi-view coding transmission in the first and second embodiments are repeated, and the description is omitted here.
- the third embodiment corresponds to the content in the first embodiment and the second embodiment, and can be obtained from the above-mentioned first embodiment and second embodiment respectively.
- the point cloud projection transmission method based on the user's perspective includes the following steps:
- edge segment Considering the spatial connection between the edge segment and the main segment, maintain the projection vector of the edge segment or change it based on the user's perspective;
- the point cloud segments that need to be observed are allocated with priority resources for encoding transmission, and the point cloud segments with a perspective that do not need to be observed are allocated for encoding and transmission compared to the priority allocation of secondary resources.
- the point cloud projection transmission system based on the user's perspective as the sending end, the point cloud receiving processing method based on the user's perspective as the receiving end, and the point cloud projection and transmission system based on the user's perspective as the receiving end can be determined by the point cloud projection transmission system based on the user's perspective.
- the cloud projection transmission method can be deduced by analogy, and also includes the content of the combination of the above-mentioned first embodiment and the second embodiment respectively, and the overlapping content of the above-mentioned first embodiment and the second embodiment is omitted here.
- the present invention corresponds to the above-mentioned first embodiment, and also provides a point cloud projection transmission method based on the user's perspective, which is characterized in that it includes: preliminarily dividing the point cloud object on the standard vector to obtain the point cloud segment to be processed; The point cloud fragments to be processed are further divided to obtain the point cloud fragments of the corresponding perspective category, including the point cloud fragments that need to be observed and the point cloud fragments that do not need to be observed; the point cloud fragments that need to be observed are assigned priority resources for encoding transmission, and The point cloud segment of the perspective that does not need to be observed is preferentially allocated to secondary resources for encoding transmission.
- the point cloud projection transmission system based on the user's perspective as the sending end, the point cloud receiving processing method based on the user's perspective as the receiving end, and the point cloud projection and transmission system based on the user's perspective as the receiving end can be determined by the point cloud projection transmission system based on the user's perspective.
- the cloud projection transmission method can be deduced by analogy, and it also includes the content of the above-mentioned first embodiment, and the overlapping content of the above-mentioned first embodiment is omitted here.
- the present invention corresponds to the above-mentioned second embodiment, and also provides a point cloud projection transmission method based on a user's perspective, which is characterized in that it includes: a point cloud segment to be processed, distinguishing a main segment and a plurality of edge segments according to the user's perspective; Considering the spatial connection between the edge segment and the main segment, the projection vector of the edge segment is maintained or changed based on the user's perspective; subsequent point cloud processing is performed after projection.
- the point cloud projection transmission system based on the user's perspective as the sending end, the point cloud receiving processing method based on the user's perspective as the receiving end, and the point cloud projection and transmission system based on the user's perspective as the receiving end can be determined by the point cloud projection transmission system based on the user's perspective.
- the cloud projection transmission method can be deduced by analogy, and it also includes the content of the above-mentioned second embodiment, and the overlapping content of the above-mentioned second embodiment is omitted here.
- the program can be stored in a computer readable storage medium, and the program can be stored in a computer readable storage medium. During execution, it may include the procedures of the above-mentioned method embodiments.
- the storage medium may be a magnetic disk, an optical disc, a read-only memory (Read-Only Memory, ROM), or a random access memory (Random Access Memory, RAM), etc.
- the embodiment of the present invention provides a computer-readable storage medium on which a computer program is stored.
- the program When the program is executed by a processor, it realizes the point cloud projection processing method based on the user's perspective as provided by the embodiment of the present invention.
- Point cloud receiving and processing method When the program is executed by a processor, it realizes the point cloud projection processing method based on the user's perspective as provided by the embodiment of the present invention. Point cloud receiving and processing method.
- the computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium.
- the computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or a combination of any of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), Erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination of the above.
- the computer-readable storage medium can be any tangible medium that contains or stores a program, and the program can be used by or in combination with an instruction execution system, apparatus, or device.
- the computer-readable signal medium may include a data signal propagated in baseband or as a part of a carrier wave, and computer-readable program code is carried therein. This propagated data signal can take many forms, including, but not limited to, electromagnetic signals, optical signals, or any suitable combination of the foregoing.
- the computer-readable signal medium may also be any computer-readable medium other than the computer-readable storage medium.
- the computer-readable medium may send, propagate, or transmit the program for use by or in combination with the instruction execution system, apparatus, or device .
- the program code contained on the computer-readable medium can be transmitted by any suitable medium, including but not limited to wireless, wire, optical cable, RF, etc., or any suitable combination of the above.
- the computer program code used to perform the operations of the present invention can be written in one or more programming languages or a combination thereof.
- the programming languages include object-oriented programming languages—such as Java, Smalltalk, C++, and also conventional Procedural programming language-such as "C" language or similar programming language.
- the program code can be executed entirely on the user's computer, partly on the user's computer, executed as an independent software package, partly on the user's computer and partly executed on a remote computer, or entirely executed on the remote computer or server.
- the remote computer can be connected to the user's computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (for example, using an Internet service provider to pass Internet connection).
- LAN local area network
- WAN wide area network
Abstract
Description
参数IndexParameter Index | 标准投影向量PlaneStandard Projection Vector Plane |
00 | (1,0,0)(1, 0, 0) |
11 | (-1,0,0)(-1, 0, 0) |
22 | (0,1,0)(0, 1, 0) |
33 | (0,-1,0)(0, -1, 0) |
44 | (0,0,1)(0, 0, 1) |
55 | (0,0,-1)(0, 0, -1) |
索引参数值(Index)Index parameter value (Index) | 投影向量面NormalProjection vector surface Normal |
00 | (1,0,0)(1, 0, 0) |
11 | (-1,0,0)(-1, 0, 0) |
22 | (0,1,0)(0, 1, 0) |
33 | (1,-1,0)(1, -1, 0) |
44 | (0,0,1)(0, 0, 1) |
55 | (0,0,-1)(0, 0, -1) |
类型Types of | 原索引参数值Original index parameter value | 原投影向量Original projection vector | 与主片段空间关联Associate with the main segment space | 优化调整后索引参数值Optimized index parameter value after adjustment | 优化调整后投影向量Optimize the adjusted projection vector |
主片段Main fragment | 00 | (1,0,0)(1, 0, 0) | ———— | 00 | (1,0,0)(1, 0, 0) |
无关片段Irrelevant fragments | 11 | (-1,0,0)(-1, 0, 0) | 保持maintain | 11 | (-1,0,0)(-1, 0, 0) |
第1边缘片段1st edge segment | 22 | (0,1,0)(0, 1, 0) | 大Big | 00 | (1,0,0)(1, 0, 0) |
第2边缘片段2nd edge segment | 22 | (0,1,0)(0, 1, 0) | 不大Not big | 22 | (0,1,0)(0, 1, 0) |
第3边缘片段3rd edge segment | 44 | (0,0,1)(0, 0, 1) | 大Big | 00 | (1,0,0)(1, 0, 0) |
第4边缘片段4th edge segment | 55 | (0,0,-1)(0, 0, -1) | 不大Not big | 55 | (0,0,-1)(0, 0, -1) |
第5边缘片段5th edge segment | 33 | (0,-1,0)(0, -1, 0) | 大Big | 00 | (1,0,0)(1, 0, 0) |
第k边缘片段K-th edge segment | …… | …… | …… | …… | …… |
Claims (30)
- 一种基于用户视角的点云投影传输方法,其特征在于,包括:A point cloud projection transmission method based on the user's perspective, which is characterized in that it includes:将点云对象在标准向量上初步划分,得到待处理点云片段;Preliminarily divide the point cloud object on the standard vector to obtain the point cloud segment to be processed;基于用户视角确定投影传输方案,对待处理点云片段进行处理。The projection transmission scheme is determined based on the user's perspective, and the point cloud segment to be processed is processed.
- 根据权利要求1所述的基于用户视角的点云投影传输方法,其特征在于:The point cloud projection transmission method based on the user's perspective according to claim 1, characterized in that:取距离最近的标准法向量为点云对象中点的划分类别;Take the nearest standard normal vector as the division category of the midpoint of the point cloud object;基于六个标准法向量将点云对象中点分别划分为多个一级点云片段;Divide the midpoint of the point cloud object into multiple first-level point cloud fragments based on six standard normal vectors;将同一一级点云片段中,同一法向量且距离小于预定值的点再划分为二级点云片段作为待处理云片段。In the same first-level point cloud segment, points with the same normal vector and a distance less than a predetermined value are further divided into second-level point cloud segments as cloud segments to be processed.
- 根据权利要求1所述的基于用户视角的点云投影传输方法,其特征在于,其中,投影传输方案包括:The point cloud projection transmission method based on the user's perspective according to claim 1, wherein the projection transmission scheme comprises:将待处理点云片段进一步划分得到对应视角类别的点云片段,包含需要观察到的点云片段和不需要观察到的点云片段;The point cloud fragments to be processed are further divided to obtain the point cloud fragments of the corresponding viewing angle category, including the point cloud fragments that need to be observed and the point cloud fragments that do not need to be observed;对需要观察到的点云片段分配优先资源进行编码传输,对不需要被观察得到的视角的点云片段相较于优先分配次级资源进行编码传输。The point cloud segments that need to be observed are allocated with priority resources for encoding transmission, and the point cloud segments with a perspective that do not need to be observed are allocated for encoding and transmission compared to the priority allocation of secondary resources.
- 根据权利要求1所述的基于用户视角的点云投影传输方法,其特征在于,所述投影传输方案包括以下任一种或任多种:The point cloud projection transmission method based on the user's perspective according to claim 1, wherein the projection transmission scheme includes any one or more of the following:不修改点云投影方案、改进点云投影方案、根据用户视角来修改标准法向量方案。Do not modify the point cloud projection scheme, improve the point cloud projection scheme, and modify the standard normal vector scheme according to the user's perspective.
- 根据权利要求4所述的基于用户视角的点云投影传输方法,其特征在于,所述不修改点云投影方法的方案中,包括:The point cloud projection transmission method based on the user's perspective according to claim 4, wherein the solution of not modifying the point cloud projection method comprises:将点云对象在标准向量上初步划分,通过参数列表与作为投影向量的标准向量相关联,得到待处理点云片段,The point cloud object is preliminarily divided on the standard vector, and the parameter list is associated with the standard vector as the projection vector to obtain the point cloud fragment to be processed,考虑用户视角和投影向量的方向,将待处理点云片段分为需要观察到的点云片段和不需要观察到的点云片段。Considering the user's perspective and the direction of the projection vector, the point cloud segment to be processed is divided into the point cloud segment that needs to be observed and the point cloud segment that does not need to be observed.
- 根据权利要求5所述的基于用户视角的点云投影传输方法,其特征在于,The point cloud projection transmission method based on the user's perspective according to claim 5, characterized in that,对需要观察到的点云片段和不需要观察到的点云片段,设置相对应的编解码参数,Set the corresponding codec parameters for the point cloud segments that need to be observed and the point cloud segments that do not need to be observed,记录每个点云片段的编解码参数,作为新增附加信息用于编码传输。Record the coding and decoding parameters of each point cloud segment, as new additional information for coding transmission.
- 根据权利要求4所述的基于用户视角的点云投影传输方法,其特征在于,所述改进点云投影方案中,包括:The point cloud projection transmission method based on the user's perspective according to claim 4, wherein the improved point cloud projection solution includes:将待处理点云片段进一步划分得到对应视角类别的点云片段,包含需要观察到的点云片段和不需要观察到的点云片段;The point cloud fragments to be processed are further divided to obtain the point cloud fragments of the corresponding viewing angle category, including the point cloud fragments that need to be observed and the point cloud fragments that do not need to be observed;将需要观察到的点云片段和不需要观察到的点云片段,投影到投影平面上,根据需要更改投影平面的尺寸,得到需要观察图片和不需要观察图片。Project the point cloud fragments that need to be observed and the point cloud fragments that do not need to be observed on the projection plane, and change the size of the projection plane as needed to obtain the pictures that need to be observed and the pictures that do not need to be observed.
- 根据权利要求7所述的基于用户视角的点云投影传输方法,其特征在于,进一步包括:The point cloud projection transmission method based on the user's perspective according to claim 7, characterized in that it further comprises:作为第一改进点云投影方案,在需要观察到的点云片段和不需要观察到的点云片段之间错位前提下将不需要观察图片拼合入需要观察图片中,得到一张投影图片,进行压缩编码。As the first improved point cloud projection solution, under the premise of dislocation between the point cloud fragments that need to be observed and the point cloud fragments that do not need to be observed, the pictures that do not need to be observed are merged into the pictures that need to be observed, and a projected picture is obtained. Compression coding.
- 根据权利要求8所述的基于用户视角的点云投影传输方法,其特征在于,所得到的一张投影图片形成一组视频序列,The point cloud projection transmission method based on the user's perspective according to claim 8, wherein the obtained one projection picture forms a group of video sequences,记录不需要观察到的点云片段对应的索引参数index以及尺寸更改的比例,作为新增附加信息用于编码传输。Record the index parameter index corresponding to the point cloud segment that does not need to be observed and the scale of the size change, as new additional information for encoding transmission.
- 根据权利要求7所述的基于用户视角的点云投影传输方法,其特征在于,进一步包括:The point cloud projection transmission method based on the user's perspective according to claim 7, characterized in that it further comprises:作为第二改进点云投影方案,将需要观察图片和不需要观察图片分别进行压缩编码。As the second improved point cloud projection solution, the need to observe pictures and the unneeded pictures are compressed and coded separately.
- 根据权利要求10所述的基于用户视角的点云投影传输方法,其特征在 于,The point cloud projection transmission method based on the user's perspective according to claim 10, characterized in that:投影到的需要观察图片和不需要观察图片分别形成两组视频序列,编码两组视频序列,The projected pictures that need to be observed and the pictures that do not need to be observed form two sets of video sequences respectively, and the two sets of video sequences are encoded,记录不需要观察到的点云片段对应的视频序列的更改比例,作为新增附加信息进行编码传输,进一步记录各组视频序列在编码后的码流中位置。Record the change ratio of the video sequence corresponding to the point cloud segment that does not need to be observed, and encode it for transmission as newly added additional information, and further record the position of each group of video sequences in the coded stream.
- 根据权利要求4所述的基于用户视角的点云投影传输方法,其特征在于,第三改进点云投影方案,包括:The point cloud projection transmission method based on the user's perspective according to claim 4, wherein the third improved point cloud projection solution includes:将待处理点云片段依照六个标准法向量分组,分别投影到六个投影平面上,对六个投影平面基于用户视角分别根据需求更改尺寸,分别进行压缩编码。The point cloud fragments to be processed are grouped according to six standard normal vectors and projected onto six projection planes respectively. Based on the user's perspective, the size of the six projection planes is changed according to requirements, and compression coding is performed respectively.
- 根据权利要求12所述的基于用户视角的点云投影传输方法,其特征在于,The point cloud projection transmission method based on the user's perspective according to claim 12, characterized in that,将各个标准法向量对应的点云片段分别投影到各组视频序列,编码六组视频序列,Project the point cloud segments corresponding to each standard normal vector to each group of video sequences, and encode six groups of video sequences,记录不需要观察到的点云片段对应的视频序列的更改比例,作为新的附加信息进行编码传输,进一步记录各组视频序列在编码后的码流中的位置。Record the change ratio of the video sequence corresponding to the point cloud segment that does not need to be observed, and encode and transmit it as new additional information, and further record the position of each group of video sequences in the coded stream.
- 根据权利要求4所述的基于用户视角的点云投影传输方法,其特征在于,所述根据用户视角来修改标准法向量方案中,包括:The point cloud projection transmission method based on the user's perspective according to claim 4, wherein the scheme for modifying the standard normal vector according to the user's perspective includes:根据用户观测视角确定视角向量;Determine the angle of view vector according to the user's observation angle;将该视角向量与其中一个标准法向量进行旋转映射,然后将其他的标准法向量进行同样的旋转映射处理,在总体修改量小于预定值的前提下,确定旋转映射后的投影方向。Perform rotation mapping on the view vector and one of the standard normal vectors, and then perform the same rotation mapping processing on the other standard normal vectors, and determine the projection direction after the rotation mapping on the premise that the overall modification amount is less than a predetermined value.
- 根据权利要求14所述的基于用户视角的点云投影传输方法,其特征在于:The point cloud projection transmission method based on the user's perspective according to claim 14, characterized in that:其中,将总体修改量小于预定值的步骤包括:在视角向量与标准法向量对应,并进行旋转之后,使包括视角向量在内的旋转映射后的六个标准法向量与对应原有的六个标准法向量的点乘积的和在预定范围之内取最大值或 较大值。Among them, the step of making the overall modification amount less than a predetermined value includes: after the view angle vector corresponds to the standard normal vector and rotates, the six standard normal vectors after the rotation mapping including the view angle vector correspond to the original six The sum of the dot products of the standard normal vector takes the maximum value or a larger value within a predetermined range.
- 根据权利要求15所述的基于用户视角的点云投影传输方法,其特征在于,The point cloud projection transmission method based on the user's perspective according to claim 15, characterized in that,将旋转映射调整后的标准法向量,将作为新增的附加信息进行编码传输。The standard normal vector after the rotation mapping is adjusted will be encoded and transmitted as newly added additional information.
- 根据权利要求1所述的基于用户视角的点云投影传输方法,其特征在于,包括:The point cloud projection transmission method based on the user's perspective according to claim 1, characterized in that it comprises:其中,投影传输方案包括:Among them, the projection transmission scheme includes:对待处理点云片段,依据用户视角区分出主片段和多个边缘片段;For the point cloud segment to be processed, distinguish the main segment and multiple edge segments according to the user's perspective;考虑到边缘片段和主片段之间的空间联系,对边缘片段的投影向量进行保持或基于用户视角进行变更;Considering the spatial connection between the edge segment and the main segment, maintain the projection vector of the edge segment or change it based on the user's perspective;投影后进行后续点云处理。After projection, the subsequent point cloud processing is performed.
- 根据权利要求17所述的基于用户视角的点云投影传输方法,其特征在于:The point cloud projection transmission method based on the user's perspective according to claim 17, characterized in that:其中,考虑到边缘片段和主片段之间的空间联系的判断依据,包括:Among them, taking into account the judgment basis of the spatial connection between the edge segment and the main segment, including:依据用户视角在标准法向量中确定主向量;Determine the main vector in the standard normal vector according to the user's perspective;针对主片段和边缘片段分别沿着主向量方向投影得到主二维投影图、以及边缘二维投影图;The main segment and the edge segment are respectively projected along the main vector direction to obtain the main two-dimensional projection map and the edge two-dimensional projection map;基于每个边缘二维投影图,分别与各个主二维投影图计算得到交并比;Based on the two-dimensional projection map of each edge, calculate the intersection and ratio with each main two-dimensional projection map respectively;多个交并比与预设门限值进行比较作为空间联系的判断依据。The multiple intersection ratios are compared with the preset threshold value as the basis for judging the spatial connection.
- 根据权利要求18所述的基于用户视角的点云投影传输方法,其特征在于:The point cloud projection transmission method based on the user's perspective according to claim 18, characterized in that:其中,多个交并比中存在大于预设门限值的情况下,将边缘片段的投影向量变更为主向量;多个交并比中均不存在大于预设门限值的情况下,保持该边缘片段的投影向量。Among them, if there are multiple intersection ratios greater than the preset threshold, change the projection vector of the edge segment to the main vector; if none of the multiple intersection ratios is greater than the preset threshold, keep The projection vector of this edge segment.
- 根据权利要求19所述的基于用户视角的点云投影传输方法,其特征在于:The point cloud projection transmission method based on the user's perspective according to claim 19, characterized in that:其中,预设门限值取决于点云对象的表面特性,区分物体点云对象的表面特性和人物点云对象的表面特性,Among them, the preset threshold value depends on the surface characteristics of the point cloud object, distinguishing the surface characteristics of the object point cloud object and the surface characteristics of the character point cloud object.门限值的确定可进一步依据实验结果进行调整确定。The determination of the threshold value can be further adjusted and determined based on the experimental results.
- 根据权利要求17所述的基于用户视角的点云投影传输方法,其特征在于,The point cloud projection transmission method based on the user's perspective according to claim 17, characterized in that,根据用户观看点云对象时所处的位置,用一个从点云指向用户观察点的向量表示用户的观察视角,称为视角向量,According to the position of the user when viewing the point cloud object, a vector from the point cloud to the user’s observation point is used to represent the user’s observation perspective, which is called the perspective vector.计算视角向量与六个标准法向量的距离,与视角向量距离最近的标准法向量称为该观察视角下的主向量。Calculate the distance between the viewing angle vector and the six standard normal vectors, and the standard normal vector with the closest distance to the viewing angle vector is called the principal vector under the viewing angle.
- 根据权利要求17所述的基于用户视角的点云投影传输方法,其特征在于,The point cloud projection transmission method based on the user's perspective according to claim 17, characterized in that,除了依据用户视角区分出点云对象的主片段和多个边缘片段,还包括区分出无关片段,In addition to distinguishing the main segment and multiple edge segments of the point cloud object according to the user’s perspective, it also includes distinguishing irrelevant segments.其中,将与用户视角距离最接近的标准法向量确定为主向量;Among them, the standard normal vector closest to the user's viewing angle is determined as the main vector;将多个标准法向量中与该主向量相同的待处理点云片段确定为主片段;Determine the point cloud segment to be processed that is the same as the main vector among the multiple standard normal vectors as the main segment;将多个标准法向量中与该主向量相反的待处理点云片段确定为无关片段;Determine the to-be-processed point cloud segment opposite to the main vector among the multiple standard normal vectors as irrelevant segments;将多个标准法向量中与该主向量相关联的待处理点云片段确定为边缘片段。The to-be-processed point cloud segment associated with the main vector among the multiple standard normal vectors is determined as an edge segment.
- 根据权利要求1所述的基于用户视角的点云投影传输方法,其特征在于:The point cloud projection transmission method based on the user's perspective according to claim 1, characterized in that:投影后进行后续点云处理中,利用投影关系标识字段标识出,点云对象中的点经过处理调整后的投影向量,包含点云片段索引值和相对应的投影向量值。In the subsequent point cloud processing after projection, the projection relationship identification field is used to identify the projection vector of the point in the point cloud object after processing and adjustment, including the point cloud fragment index value and the corresponding projection vector value.
- 一种基于用户视角的点云接收处理方法,其特征在于:A point cloud receiving and processing method based on the user's perspective, which is characterized in:利用用户视角而确定的编解码参数对接收到的码流得到视频序列,和对视频序列处理得到点云片段,以及对每个点云片段进行解码重建得到点云 对象。Use the codec parameters determined by the user's perspective to obtain a video sequence from the received bitstream, process the video sequence to obtain a point cloud segment, and decode and reconstruct each point cloud segment to obtain a point cloud object.
- 根据权利要求24所述的基于用户视角的点云接收处理方法,其特征在于,The method for receiving and processing a point cloud based on a user's perspective according to claim 24, wherein:根据编解码参数对每个点云片段进行编解码和后续重建的过程,包含以下任意一种或者几种:The process of encoding, decoding and subsequent reconstruction of each point cloud segment according to the encoding and decoding parameters includes any one or more of the following:所述编解码参数在发送端针对每个点云片段,区分需要观察到的点云片段和不需要观察被设置,进行后续的点云对象的重建;或者The coding and decoding parameters are set at the sending end for each point cloud segment to distinguish between the point cloud segment that needs to be observed and the point cloud segment that does not need to be observed, and the subsequent reconstruction of the point cloud object is performed; or需要根据所接收到的参数信息以及尺寸更改信息来对点云片段恢复为原始尺寸,再进行后续的点云对象的重建;或者It is necessary to restore the point cloud segment to its original size according to the received parameter information and size change information, and then perform subsequent point cloud object reconstruction; or根据各视频序列在码流中的位置恢复出各视频序列,根据所接收到的更改尺寸来将这些视频序列恢复为原始尺寸,并需要将多组视频序列所恢复出来的点云进行拼接,进行点云对象的重建;或者According to the position of each video sequence in the code stream, each video sequence is restored, and these video sequences are restored to the original size according to the received change size, and the point clouds recovered from multiple sets of video sequences need to be spliced. Reconstruction of point cloud objects; or需要根据发送端旋转映射调整后的标准法向量,进行点云投影的逆过程,从而从二维视频序列重建出三维点云序列,进行点云对象的重建。It is necessary to perform the inverse process of point cloud projection according to the standard normal vector adjusted by the rotation mapping of the transmitting end, so as to reconstruct a three-dimensional point cloud sequence from the two-dimensional video sequence and perform the reconstruction of the point cloud object.
- 一种基于用户视角的点云投影传输系统,其特征在于:A point cloud projection transmission system based on the user's perspective is characterized by:片段生成模块,用于将点云对象在标准向量上初步划分,得到待处理点云片段;The fragment generation module is used to preliminarily divide the point cloud object on the standard vector to obtain the point cloud fragment to be processed;投影编码处理模块,基于用户视角确定投影传输方案,对待处理点云片段进行处理。The projection coding processing module determines the projection transmission scheme based on the user's perspective, and processes the point cloud segment to be processed.
- 一种基于用户视角的点云接收处理系统,其特征在于:A point cloud receiving and processing system based on the user's perspective is characterized by:逆处理模块,利用用户视角而确定的编解码参数对接收到的码流得到视频序列,和对视频序列处理得到点云片段,以及对每个点云片段进行解码重建得到点云对象。The inverse processing module uses the codec parameters determined from the user's perspective to obtain a video sequence from the received bitstream, process the video sequence to obtain a point cloud segment, and decode and reconstruct each point cloud segment to obtain a point cloud object.
- 一种基于用户视角的点云投影传输方法,其特征在于,包括:A point cloud projection transmission method based on the user's perspective, which is characterized in that it includes:将点云对象在标准向量上初步划分,得到待处理点云片段;Preliminarily divide the point cloud object on the standard vector to obtain the point cloud segment to be processed;将待处理点云片段进一步划分得到对应视角类别的点云片段,包含需要观 察到的点云片段和不需要观察到的点云片段;The point cloud fragments to be processed are further divided to obtain the point cloud fragments of the corresponding viewing angle category, including the point cloud fragments that need to be observed and the point cloud fragments that do not need to be observed;对需要观察到的点云片段分配优先资源进行编码传输,对不需要被观察得到的视角的点云片段相较于优先分配次级资源进行编码传输The point cloud fragments that need to be observed are allocated with priority resources for encoding transmission, and the point cloud fragments that do not need to be observed are compared with the priority allocation of secondary resources for encoding transmission.
- 一种基于用户视角的点云投影传输方法,其特征在于,包括:A point cloud projection transmission method based on the user's perspective, which is characterized in that it includes:对待处理点云片段,依据用户视角区分出主片段和多个边缘片段;For the point cloud segment to be processed, distinguish the main segment and multiple edge segments according to the user's perspective;考虑到边缘片段和主片段之间的空间联系,对边缘片段的投影向量进行保持或基于用户视角进行变更;Considering the spatial connection between the edge segment and the main segment, maintain the projection vector of the edge segment or change it based on the user's perspective;投影后进行后续点云处理。After projection, the subsequent point cloud processing is performed.
- 一种基于用户视角的点云投影传输方法,其特征在于,包括:A point cloud projection transmission method based on the user's perspective, which is characterized in that it includes:将点云对象在标准向量上初步划分,得到待处理点云片段,依据用户视角区分出主片段和多个边缘片段;Preliminarily divide the point cloud object on the standard vector to obtain the point cloud segment to be processed, and distinguish the main segment and multiple edge segments according to the user's perspective;考虑到边缘片段和主片段之间的空间联系,对边缘片段的投影向量进行保持或基于用户视角进行变更;Considering the spatial connection between the edge segment and the main segment, maintain the projection vector of the edge segment or change it based on the user's perspective;将调整变更后的主片段作为需要观察到的点云片段,调整变更后的边缘片段和无关片段作为不需要观察到的点云片段;以及Take the adjusted and changed main segment as the point cloud segment that needs to be observed, and adjust and change the edge segment and irrelevant segments as the point cloud segment that does not need to be observed; and对需要观察到的点云片段分配优先资源进行编码传输,对不需要被观察得到的视角的点云片段相较于优先分配次级资源进行编码传输。The point cloud segments that need to be observed are allocated with priority resources for encoding transmission, and the point cloud segments with a perspective that do not need to be observed are allocated for encoding and transmission compared to the priority allocation of secondary resources.
Applications Claiming Priority (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910827750.9A CN112449171B (en) | 2019-09-03 | 2019-09-03 | Encoding method, system and medium for point cloud view-division transmission |
CN201910827750.9 | 2019-09-03 | ||
CN202010863785.0 | 2020-08-25 | ||
CN202010863785.0A CN114119346A (en) | 2020-08-25 | 2020-08-25 | Point cloud projection processing method and system based on user view angle and receiving processing method |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2021043136A1 true WO2021043136A1 (en) | 2021-03-11 |
Family
ID=74852153
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2020/112857 WO2021043136A1 (en) | 2019-09-03 | 2020-09-01 | Point cloud projection transmission method, and reception processing method and system |
Country Status (1)
Country | Link |
---|---|
WO (1) | WO2021043136A1 (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20190087979A1 (en) * | 2017-09-18 | 2019-03-21 | Apple Inc. | Point cloud compression |
CN109523581A (en) * | 2017-09-19 | 2019-03-26 | 华为技术有限公司 | A kind of method and apparatus of three-dimensional point cloud alignment |
CN110012279A (en) * | 2018-01-05 | 2019-07-12 | 上海交通大学 | Divide visual angle compression and transmission method and system based on 3D point cloud data |
CN110070613A (en) * | 2019-04-26 | 2019-07-30 | 东北大学 | Large-scale three dimensional scene web page display method based on model compression and asynchronous load |
CN110070608A (en) * | 2019-04-11 | 2019-07-30 | 浙江工业大学 | A method of being automatically deleted the three-dimensional reconstruction redundant points based on image |
-
2020
- 2020-09-01 WO PCT/CN2020/112857 patent/WO2021043136A1/en active Application Filing
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20190087979A1 (en) * | 2017-09-18 | 2019-03-21 | Apple Inc. | Point cloud compression |
CN109523581A (en) * | 2017-09-19 | 2019-03-26 | 华为技术有限公司 | A kind of method and apparatus of three-dimensional point cloud alignment |
CN110012279A (en) * | 2018-01-05 | 2019-07-12 | 上海交通大学 | Divide visual angle compression and transmission method and system based on 3D point cloud data |
CN110070608A (en) * | 2019-04-11 | 2019-07-30 | 浙江工业大学 | A method of being automatically deleted the three-dimensional reconstruction redundant points based on image |
CN110070613A (en) * | 2019-04-26 | 2019-07-30 | 东北大学 | Large-scale three dimensional scene web page display method based on model compression and asynchronous load |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20210174551A1 (en) | Mesh compression via point cloud representation | |
US11010955B2 (en) | Point cloud mapping | |
US20220116659A1 (en) | A method, an apparatus and a computer program product for volumetric video | |
CN117768653A (en) | Method and apparatus for encoding and decoding volumetric video | |
CN108027983A (en) | For the generation of the triangular mesh of 3-D view | |
US20210006806A1 (en) | An apparatus, a method and a computer program for volumetric video | |
CN107454468A (en) | The method, apparatus and stream being formatted to immersion video | |
CN103828359A (en) | Representation and coding of multi-view images using tapestry encoding | |
JP2023514853A (en) | projection-based mesh compression | |
KR101933037B1 (en) | Apparatus for reproducing 360 degrees video images for virtual reality | |
CN112449171B (en) | Encoding method, system and medium for point cloud view-division transmission | |
US20200236401A1 (en) | Point cloud coding using homography transform | |
JP7344988B2 (en) | Methods, apparatus, and computer program products for volumetric video encoding and decoding | |
Zhu et al. | An improved depth image based virtual view synthesis method for interactive 3D video | |
KR102491674B1 (en) | Method and apparatus for generating virtual viewpoint image | |
CN103905812A (en) | Texture/depth combination up-sampling method | |
US20210092345A1 (en) | Unified coding of 3d objects and scenes | |
WO2021043136A1 (en) | Point cloud projection transmission method, and reception processing method and system | |
WO2023272510A1 (en) | Methods and apparatuses for generating, performing data processing on, coding and decoding multiplane image | |
Lin et al. | Fast multi-view image rendering method based on reverse search for matching | |
EP4133719A1 (en) | A method, an apparatus and a computer program product for volumetric video coding | |
JP2023541271A (en) | High density mesh compression | |
WO2019185983A1 (en) | A method, an apparatus and a computer program product for encoding and decoding digital volumetric video | |
CN116723305B (en) | Virtual viewpoint quality enhancement method based on generation type countermeasure network | |
US20230306683A1 (en) | Mesh patch sub-division |
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: 20860252 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: 20860252 Country of ref document: EP Kind code of ref document: A1 |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 20860252 Country of ref document: EP Kind code of ref document: A1 |
|
32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 19-05-2023) |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 20860252 Country of ref document: EP Kind code of ref document: A1 |