Back to results
US 12,614,314 B2Granted
Method and apparatus of encoding/decoding point cloud geometry data sensed by at least one sensor
AssigneeBeijing Xiaomi Mobile Software Co., Ltd.
InventorsSebastien Lasserre, Jonathan Taquet
Life of this patent
Estimating the term…
Abstract
A method and apparatus for encoding/decoding a point cloud may use any type of sensor following a sensing path. The method obtains coarse representations of sensed points and encodes the sensing path and the coarse representations. The sensing path and coarse representations of points are decoded, and points of the point cloud are reconstructed from the decoded sensing path and the decoded coarse representations.
Claims20
A method of encoding points of a point cloud sensed by at least one sensor associated with a sensor index into a bitstream of encoded point cloud data, wherein the method comprises:
- encoding (120), into the bitstream, an information (I(SP)) representing a sensing path, wherein
- the sensing path is defined from sensor characteristics in a two-dimensional angular coordinate space comprising an azimuthal coordinate representative of an azimuthal angle representing a sensing angle of a sensor relative to a referential and an elevation coordinate representative of an elevation angle of a sensor relative to a horizontal referential plane;
- the sensing path is used to sense the points of the point cloud according to ordered coarse points representing potential locations of sensed points of the point cloud, each coarse point being defined in the two-dimensional angular coordinate space from one sample index associated with a sensing time instant and one sensor index associated with a sensor;
- said sample index is defined along the sensing path according to a sensing scheme;
- an order index is associated with each coarse point according to a rank of said coarse point among the ordered coarse points;
- for each point (P) of the point cloud,
- obtaining (130) one sample index(s) associated with a sensing time instant of the point (P) of the point cloud along the sensing path and a sensor index (λ) of a sensor that sensed the point (P) of the point cloud;
- encoding (140), into the bitstream, one order index difference (Ao) representing a difference between an order index (o(P)) of a coarse point identified by the sample index and the sensor index, and an order index (o(P−1)) of another coarse point associated with a previous sensed point (P−1) of the point cloud;
- obtaining (150) one sensing point (Pover(s)) in the two-dimensional angular coordinate space from the sample index(s) and the sensing scheme;
- obtaining (160) one coarse point (Pλ(s)) in the two-dimensional angular coordinate space from the sensing point (Pover(s)), the sample index(s) and the sensor index (λ);
- and
- encoding (170), into the bitstream, the point (P) of the point cloud relatively to the coarse point (Pλ(s)).
9 more elements
A method of decoding points of a point cloud sensed by at least one sensor associated with a sensor index from a bitstream of encoded point cloud data, wherein the method comprises:
- decoding (210), from the bitstream, an information (I(SP)) representing a sensing path wherein
- the sensing path is defined in a two-dimensional angular coordinate space comprising an azimuthal coordinate representative of an azimuthal angle representing a sensing angle of a sensor relative to a referential and an elevation coordinate representative of an elevation angle of a sensor relative to a horizontal referential plane;
- the sensing path has been used to sense the points of the point cloud according to ordered coarse points representing potential locations of sensed points of the point cloud, each coarse point being defined in the two-dimensional angular coordinate space from one sample index associated with a sensing time instant and one sensor index associated with a sensor, said one sample index being defined along the sensing path according to a sensing scheme;
- an order index is associated with each coarse point according to a rank of said coarse point among the ordered coarse points;
- for each point (P) of the point cloud,
- decoding (220), from the bitstream, one order index difference (Δo) representing a difference between an order index (o(P)) of a coarse point associated with the point (P) of the point cloud, and an order index (o(P−1)) of another coarse point associated with a previous decoded point (P−1) of the point cloud;
- obtaining (230) one sensor index (λ) associated with a sensor that sensed the point (P) of the point cloud and one sample index(s) associated with the sensing time instant of the point (P) of the point cloud along the sensing path from an order index (o(P)) obtained from the decoded order index difference (Δo);
- obtaining (150) one sensing point (Pover(s)) in the two-dimensional angular coordinate space from the sample index(s) and the sensing scheme;
- obtaining (160) a coarse point (Pλ(s)) in the two-dimensional angular coordinate space from the sensing point (Pover(s)), the sample index(s) and the sensor index (λ);
- and
- decoding (240) a point of the point cloud relatively to the coarse point (Pλ(s)).
8 more elements
An apparatus of encoding points of a point cloud sensed by at least one sensor associated with a sensor index into a bitstream of encoded point cloud data, wherein the apparatus comprises at least one processor configured to:
- encode, into the bitstream, an information (I(SP)) representing a sensing path, wherein
- the sensing path is defined from sensor characteristics in a two-dimensional angular coordinate space comprising an azimuthal coordinate representative of an azimuthal angle representing a sensing angle of a sensor relative to a referential and an elevation coordinate representative of an elevation angle of a sensor relative to a horizontal referential plane;
- the sensing path is used to sense the points of the point cloud according to ordered coarse points representing potential locations of sensed points of the point cloud, each coarse point being defined in the two-dimensional angular coordinate space from one sample index associated with a sensing time instant and one sensor index associated with a sensor;
- said sample index is defined along the sensing path according to a sensing scheme;
- an order index is associated with each coarse point according to a rank of said coarse point among the ordered coarse points;
- for each point (P) of the point cloud,
- obtain one sample index(s) associated with a sensing time instant of the point (P) of the point cloud along the sensing path and a sensor index (λ) of a sensor that sensed the point (P) of the point cloud;
- encode, into the bitstream, one order index difference (Δo) representing a difference between an order index (o(P)) of a coarse point identified by the sample index and the sensor index, and an order index (o(P−1)) of another coarse point associated with a previous sensed point (P−1) of the point cloud;
- obtain one sensing point (Pover(s)) in the two-dimensional angular coordinate space from the sample index(s) and the sensing scheme;
- obtain one coarse point (Pλ(s)) in the two-dimensional angular coordinate space from the sensing point (Pover(s)), the sample index(s) and the sensor index (λ);
- and
- encode, into the bitstream, the point (P) of the point cloud relatively to the coarse point (Pλ(s)).
9 more elements
Dependencies are inferred from the claim text.