概览硬件软件云平台
OAK 应用
DepthAI
AI
软件栈
故障排除

本页目录

  • 使用说明
  • 参考

PointCloudData

PointCloudData 封装了3D空间信息，表示3D空间中点的集合。点云中的每个点都有其自身的位置（X、Y、Z坐标）。PointCloudData 用于表示 PointCloud 节点的输出，并可用于执行各种空间分析和重建任务。Setter方法仅用于向 PointCloudData 消息提供元数据（将用于点云的录制和回放）。

使用说明

  • 有组织与过滤 — 当在 PointCloud 节点上使用 setOrganized(true) 时，getWidth() / getHeight() 匹配源深度帧的尺寸，并包含无效点（z ≤ 0）。当过滤时（默认），height 为 1，width 等于有效点的数量。
  • 颜色数据 — 当 isColor() 为 true 时，使用 getPointsRGB() 获取位置和RGBA颜色。在 Python 中，这会返回一个元组 (points, colors)，其中 points 为 np.ndarray (N, 3) float32，colors 为 np.ndarray (N, 4) uint8。
  • 边界框 — getMinX/Y/Z() 和 getMaxX/Y/Z() 返回以配置长度单位表示的轴对齐边界。

参考

PointCloudData 的详细API提供了对3D点云数据生成、操作和检索的控制。

Python

class

dai::PointCloudData

#include PointCloudData.hpp
variable
int64_t sequenceNum
variable
variable
Timestamp tsDevice
variable
std::optional< Timestamp > tsSystem
variable
std::optional< ImgTransformation > transformation
function
PointCloudData()
Construct PointCloudData message.
function
~PointCloudData()
function
std::vector< Point3f > getPoints()
function
std::vector< Point3fRGBA > getPointsRGB()
function
void setPoints(const std::vector< Point3f > & points)
function
void setPointsRGB(const std::vector< Point3fRGBA > & points)
function
unsigned int getInstanceNum()
Retrieves instance number
function
unsigned int getWidth()
Retrieves the height in pixels - in case of a sparse point cloud, this represents the hight of the frame which was used to generate the point cloud
function
unsigned int getHeight()
Retrieves the height in pixels - in case of a sparse point cloud, this represents the hight of the frame which was used to generate the point cloud
function
float getMinX()
Retrieves minimal x coordinate in depth units (millimeter by default)
function
float getMinY()
Retrieves minimal y coordinate in depth units (millimeter by default)
function
float getMinZ()
Retrieves minimal z coordinate in depth units (millimeter by default)
function
float getMaxX()
Retrieves maximal x coordinate in depth units (millimeter by default)
function
float getMaxY()
Retrieves maximal y coordinate in depth units (millimeter by default)
function
float getMaxZ()
Retrieves maximal z coordinate in depth units (millimeter by default)
function
bool isSparse()
Retrieves whether point cloud is sparse
function
bool isOrganized()
Retrieves whether point cloud is organized (height > 1) Organized point clouds have width x height structure from the original image Sparse point clouds have height == 1 and only contain valid points
function
bool isColor()
Retrieves whether point cloud is color
function
PointCloudData & setWidth(unsigned int width)
Specifies frame width
Parameters
  • width: frame width
function
PointCloudData & setHeight(unsigned int height)
Specifies frame height
Parameters
  • height: frame height
function
PointCloudData & setSize(unsigned int width, unsigned int height)
Specifies frame size
Parameters
  • height: frame height
  • width: frame width
function
PointCloudData & setSize(std::tuple< unsigned int, unsigned int > size)
Specifies frame size
Parameters
  • size: frame size
function
PointCloudData & setMinX(float val)
Specifies minimal x coordinate in depth units (millimeter by default)
Parameters
  • val: minimal x coordinate in depth units (millimeter by default)
function
PointCloudData & setMinY(float val)
Specifies minimal y coordinate in depth units (millimeter by default)
Parameters
  • val: minimal y coordinate in depth units (millimeter by default)
function
PointCloudData & setMinZ(float val)
Specifies minimal z coordinate in depth units (millimeter by default)
Parameters
  • val: minimal z coordinate in depth units (millimeter by default)
function
PointCloudData & setMaxX(float val)
Specifies maximal x coordinate in depth units (millimeter by default)
Parameters
  • val: maximal x coordinate in depth units (millimeter by default)
function
PointCloudData & setMaxY(float val)
Specifies maximal y coordinate in depth units (millimeter by default)
Parameters
  • val: maximal y coordinate in depth units (millimeter by default)
function
PointCloudData & setMaxZ(float val)
Specifies maximal z coordinate in depth units (millimeter by default)
Parameters
  • val: maximal z coordinate in depth units (millimeter by default)
function
PointCloudData & setSparse(bool val)
Specifies whether point cloud is sparse
Parameters
  • val: whether point cloud is sparse
function
PointCloudData & setColor(bool val)
Specifies whether point cloud is color
Parameters
  • val: whether point cloud is color
function
PointCloudData & setInstanceNum(unsigned int instanceNum)
Specifies instance number
Parameters
  • instanceNum: instance number
function
const ImgTransformation & getTransformation()
Retrieves image transformation data
function
PointCloudData transformTo(const ImgTransformation & target)
Returns a copy of this point cloud.Due to efficiency reasons, point cloud remapping is currently not implemented. Point clouds should be generated from already aligned source messages instead of being transformed after the fact. Use ImageAlign on the source inputs before creating the point cloud.
Parameters
  • target: Target image transformation.
function
PointCloudData & setTransformation(const ImgTransformation & transformation)
Specifies image transformation data
Parameters
  • transformation: transformation data
function
PointCloudData & updateBoundingBox()
Recomputes the bounding box (min/max X, Y, Z) from the current point data. All stored points are included regardless of their z value. If the cloud is empty, all bounds are set to 0.
inline function
void getPclData()
inline function
void setPclData(T...)
function
void serialize(std::vector< std::uint8_t > & metadata, DatatypeEnum & datatype)
inline function
DatatypeEnum getDatatype()
function
DEPTHAI_SERIALIZE(PointCloudData, width, height, minx, miny, minz, maxx, maxy, maxz, instanceNum, color, transformation, ts, tsDevice, tsSystem, sequenceNum)
function
int64_t getSequenceNum()
Retrieves image sequence number
function
std::chrono::time_point< std::chrono::steady_clock, std::chrono::steady_clock::duration > getTimestamp()
Retrieves timestamp related to dai::Clock::now()
function
std::chrono::time_point< std::chrono::steady_clock, std::chrono::steady_clock::duration > getTimestampDevice()
Retrieves timestamp directly captured from device's monotonic clock, not synchronized to host time. Used mostly for debugging
function
std::optional< std::chrono::time_point< std::chrono::system_clock, std::chrono::system_clock::duration > > getTimestampSystem()
Retrieves timestamp directly captured from device's system clock, that can be synchronized using PTP
function
std::optional< ImgTransformation > getTransformation()
Returns the current transformation if set, else std::nullopt.
function
void setTransformation(const ImgTransformation & transformation)
Sets the current transformation.
struct

dai::PointCloudData::dependent_false

C++

class

dai::PointCloudData

#include PointCloudData.hpp
variable
int64_t sequenceNum
variable
variable
Timestamp tsDevice
variable
std::optional< Timestamp > tsSystem
variable
std::optional< ImgTransformation > transformation
function
PointCloudData()
Construct PointCloudData message.
function
~PointCloudData()
function
std::vector< Point3f > getPoints()
function
std::vector< Point3fRGBA > getPointsRGB()
function
void setPoints(const std::vector< Point3f > & points)
function
void setPointsRGB(const std::vector< Point3fRGBA > & points)
function
unsigned int getInstanceNum()
Retrieves instance number
function
unsigned int getWidth()
Retrieves the height in pixels - in case of a sparse point cloud, this represents the hight of the frame which was used to generate the point cloud
function
unsigned int getHeight()
Retrieves the height in pixels - in case of a sparse point cloud, this represents the hight of the frame which was used to generate the point cloud
function
float getMinX()
Retrieves minimal x coordinate in depth units (millimeter by default)
function
float getMinY()
Retrieves minimal y coordinate in depth units (millimeter by default)
function
float getMinZ()
Retrieves minimal z coordinate in depth units (millimeter by default)
function
float getMaxX()
Retrieves maximal x coordinate in depth units (millimeter by default)
function
float getMaxY()
Retrieves maximal y coordinate in depth units (millimeter by default)
function
float getMaxZ()
Retrieves maximal z coordinate in depth units (millimeter by default)
function
bool isSparse()
Retrieves whether point cloud is sparse
function
bool isOrganized()
Retrieves whether point cloud is organized (height > 1) Organized point clouds have width x height structure from the original image Sparse point clouds have height == 1 and only contain valid points
function
bool isColor()
Retrieves whether point cloud is color
function
PointCloudData & setWidth(unsigned int width)
Specifies frame width
Parameters
  • width: frame width
function
PointCloudData & setHeight(unsigned int height)
Specifies frame height
Parameters
  • height: frame height
function
PointCloudData & setSize(unsigned int width, unsigned int height)
Specifies frame size
Parameters
  • height: frame height
  • width: frame width
function
PointCloudData & setSize(std::tuple< unsigned int, unsigned int > size)
Specifies frame size
Parameters
  • size: frame size
function
PointCloudData & setMinX(float val)
Specifies minimal x coordinate in depth units (millimeter by default)
Parameters
  • val: minimal x coordinate in depth units (millimeter by default)
function
PointCloudData & setMinY(float val)
Specifies minimal y coordinate in depth units (millimeter by default)
Parameters
  • val: minimal y coordinate in depth units (millimeter by default)
function
PointCloudData & setMinZ(float val)
Specifies minimal z coordinate in depth units (millimeter by default)
Parameters
  • val: minimal z coordinate in depth units (millimeter by default)
function
PointCloudData & setMaxX(float val)
Specifies maximal x coordinate in depth units (millimeter by default)
Parameters
  • val: maximal x coordinate in depth units (millimeter by default)
function
PointCloudData & setMaxY(float val)
Specifies maximal y coordinate in depth units (millimeter by default)
Parameters
  • val: maximal y coordinate in depth units (millimeter by default)
function
PointCloudData & setMaxZ(float val)
Specifies maximal z coordinate in depth units (millimeter by default)
Parameters
  • val: maximal z coordinate in depth units (millimeter by default)
function
PointCloudData & setSparse(bool val)
Specifies whether point cloud is sparse
Parameters
  • val: whether point cloud is sparse
function
PointCloudData & setColor(bool val)
Specifies whether point cloud is color
Parameters
  • val: whether point cloud is color
function
PointCloudData & setInstanceNum(unsigned int instanceNum)
Specifies instance number
Parameters
  • instanceNum: instance number
function
const ImgTransformation & getTransformation()
Retrieves image transformation data
function
PointCloudData transformTo(const ImgTransformation & target)
Returns a copy of this point cloud.Due to efficiency reasons, point cloud remapping is currently not implemented. Point clouds should be generated from already aligned source messages instead of being transformed after the fact. Use ImageAlign on the source inputs before creating the point cloud.
Parameters
  • target: Target image transformation.
function
PointCloudData & setTransformation(const ImgTransformation & transformation)
Specifies image transformation data
Parameters
  • transformation: transformation data
function
PointCloudData & updateBoundingBox()
Recomputes the bounding box (min/max X, Y, Z) from the current point data. All stored points are included regardless of their z value. If the cloud is empty, all bounds are set to 0.
inline function
void getPclData()
inline function
void setPclData(T...)
function
void serialize(std::vector< std::uint8_t > & metadata, DatatypeEnum & datatype)
inline function
DatatypeEnum getDatatype()
function
DEPTHAI_SERIALIZE(PointCloudData, width, height, minx, miny, minz, maxx, maxy, maxz, instanceNum, color, transformation, ts, tsDevice, tsSystem, sequenceNum)
function
int64_t getSequenceNum()
Retrieves image sequence number
function
std::chrono::time_point< std::chrono::steady_clock, std::chrono::steady_clock::duration > getTimestamp()
Retrieves timestamp related to dai::Clock::now()
function
std::chrono::time_point< std::chrono::steady_clock, std::chrono::steady_clock::duration > getTimestampDevice()
Retrieves timestamp directly captured from device's monotonic clock, not synchronized to host time. Used mostly for debugging
function
std::optional< std::chrono::time_point< std::chrono::system_clock, std::chrono::system_clock::duration > > getTimestampSystem()
Retrieves timestamp directly captured from device's system clock, that can be synchronized using PTP
function
std::optional< ImgTransformation > getTransformation()
Returns the current transformation if set, else std::nullopt.
function
void setTransformation(const ImgTransformation & transformation)
Sets the current transformation.
struct

dai::PointCloudData::dependent_false

需要帮助？

请前往 OAKChina 官网 获取技术支持或解答您的任何疑问。