Depth Sensing Module

Classes
struct	RuntimeParameters
	Structure containing parameters that defines the behavior of sl::Camera::grab(). More...
class	CustomDepthData
	Class holding an externally computed disparity or depth map, to be ingested with sl::Camera::ingestCustomDepth(). More...
struct	VoxelMeasureParameters
	Parameters controlling voxel decimation in sl::Camera::retrieveVoxelMeasure. More...
struct	CameraParameters
	Structure containing the intrinsic parameters of a camera. More...
struct	CalibrationParameters
	Structure containing intrinsic and extrinsic parameters of the camera (translation and rotation). More...

Enumerations
enum class	DEPTH_MODE
	Lists available depth computation modes. More...
enum class	MEASURE
	Lists retrievable measures. More...
enum class	REGION_OF_INTEREST_AUTO_DETECTION_STATE
	Lists the different states of region of interest auto detection. More...
enum class	VOXELIZATION_MODE
	Controls how voxel size adapts with depth in sl::Camera::retrieveVoxelMeasure. More...
enum class	CUSTOM_DEPTH_FORMAT
	Lists the content type of the map ingested with sl::Camera::ingestCustomDepth(). More...
enum class	CUSTOM_CONFIDENCE_CONVENTION
	Lists the value convention of the confidence map optionally ingested with sl::Camera::ingestCustomDepth(). More...

Enumeration Type Documentation

DEPTH_MODE

enum DEPTH_MODE

strong

Lists available depth computation modes.

Enumerator
NONE	No depth map computation. Only rectified stereo images will be available.
PERFORMANCE	Deprecated: Computation mode optimized for speed.
QUALITY	Deprecated: Computation mode designed for challenging areas with untextured surfaces.
ULTRA	Deprecated: Computation mode that favors edges and sharpness. Requires more GPU memory and computation power.
NEURAL_LIGHT	End to End Neural disparity estimation. Requires AI module.
NEURAL	End to End Neural disparity estimation. Requires AI module.
NEURAL_PLUS	More accurate Neural disparity estimation. Requires AI module.
CUSTOM	No internal depth computation. The depth (or disparity) is provided for each frame by the user with sl::Camera::ingestCustomDepth(), between sl::Camera::read() and sl::Camera::grab(). All depth-dependent modules then run on the ingested data. Warning Experimental feature: downstream modules may not be fully robust to arbitrary custom depth/confidence input and can produce unexpected behavior or reduced accuracy. Note Positional tracking only supports sl::POSITIONAL_TRACKING_MODE::GEN_3 in this mode. Depth stabilization uses the positional tracking pose when tracking is enabled; otherwise a static camera is assumed. For a moving camera, enable positional tracking (GEN_3) or set sl::InitParameters::depth_stabilization to 0.

MEASURE

enum MEASURE

strong

Lists retrievable measures.

Enumerator
DISPARITY	Disparity map. Each pixel contains 1 float. Type: sl::MAT_TYPE::F32_C1
DEPTH	Depth map in sl::UNIT defined in sl::InitParameters.coordinate_units. Each pixel contains 1 float. Type: sl::MAT_TYPE::F32_C1.
CONFIDENCE	Certainty/confidence of the depth map. Each pixel contains 1 float. Type: sl::MAT_TYPE::F32_C1
XYZ	Point cloud. Each pixel contains 4 float (X, Y, Z, not used). Type: sl::MAT_TYPE::F32_C4
XYZRGBA	Colored point cloud. Each pixel contains 4 float (X, Y, Z, color). The color should to be read as an unsigned char[4] representing the RGBA color. Type: sl::MAT_TYPE::F32_C4
XYZBGRA	Colored point cloud. Each pixel contains 4 float (X, Y, Z, color). The color should to be read as an unsigned char[4] representing the BGRA color. Type: sl::MAT_TYPE::F32_C4
XYZARGB	Colored point cloud. Each pixel contains 4 float (X, Y, Z, color). The color should to be read as an unsigned char[4] representing the ARGB color. Type: sl::MAT_TYPE::F32_C4
XYZABGR	Colored point cloud. Each pixel contains 4 float (X, Y, Z, color). The color should to be read as an unsigned char[4] representing the ABGR color. Type: sl::MAT_TYPE::F32_C4
NORMALS	Normal vectors map. Each pixel contains 4 float (X, Y, Z, 0). Type: sl::MAT_TYPE::F32_C4.
DISPARITY_RIGHT	Disparity map for right sensor. Each pixel contains 1 float. Type: sl::MAT_TYPE::F32_C1.
DEPTH_RIGHT	Depth map for right sensor. Each pixel contains 1 float. Type: sl::MAT_TYPE::F32_C1.
XYZ_RIGHT	Point cloud for right sensor. Each pixel contains 4 float (X, Y, Z, not used). Type: sl::MAT_TYPE::F32_C4
XYZRGBA_RIGHT	Colored point cloud for right sensor. Each pixel contains 4 float (X, Y, Z, color). The color should to be read as an unsigned char[4] representing the RGBA color. Type: sl::MAT_TYPE::F32_C4
XYZBGRA_RIGHT	Colored point cloud for right sensor. Each pixel contains 4 float (X, Y, Z, color). The color should to be read as an unsigned char[4] representing the BGRA color. Type: sl::MAT_TYPE::F32_C4
XYZARGB_RIGHT	Colored point cloud for right sensor. Each pixel contains 4 float (X, Y, Z, color). The color should to be read as an unsigned char[4] representing the ARGB color. Type: sl::MAT_TYPE::F32_C4
XYZABGR_RIGHT	Colored point cloud for right sensor. Each pixel contains 4 float (X, Y, Z, color). The color should to be read as an unsigned char[4] representing the ABGR color. Type: sl::MAT_TYPE::F32_C4
NORMALS_RIGHT	Normal vectors map for right view. Each pixel contains 4 float (X, Y, Z, 0). Type: sl::MAT_TYPE::F32_C4.
DEPTH_U16_MM	Depth map in millimeter whatever the sl::UNIT defined in sl::InitParameters.coordinate_units. Invalid values are set to 0 and depth values are clamped at 65000. Each pixel contains 1 unsigned short. Type: sl::MAT_TYPE::U16_C1.
DEPTH_U16_MM_RIGHT	Depth map in millimeter for right sensor. Each pixel contains 1 unsigned short. Type: sl::MAT_TYPE::U16_C1.

REGION_OF_INTEREST_AUTO_DETECTION_STATE

enum REGION_OF_INTEREST_AUTO_DETECTION_STATE

strong

Lists the different states of region of interest auto detection.

Enumerator
RUNNING	The region of interest auto detection is initializing.
READY	The region of interest mask is ready, if auto_apply was enabled, the region of interest mask is being used
NOT_ENABLED	The region of interest auto detection is not enabled

VOXELIZATION_MODE

enum VOXELIZATION_MODE

strong

Controls how voxel size adapts with depth in sl::Camera::retrieveVoxelMeasure.

Enumerator
FIXED	No adaptation. VoxelMeasureParameters::voxel_size is used uniformly everywhere.
STEREO_UNCERTAINTY	Quadratic growth matching stereo depth noise: Z²/(f·B), scaled by VoxelMeasureParameters::resolution_scale . Voxels grow larger at distance where stereo depth precision is lower.
LINEAR	Linear growth with depth Z, scaled by VoxelMeasureParameters::resolution_scale . Suits sensors with linearly increasing noise (e.g. lidar, ToF).

CUSTOM_DEPTH_FORMAT

enum CUSTOM_DEPTH_FORMAT

strong

Lists the content type of the map ingested with sl::Camera::ingestCustomDepth().

Enumerator
DISPARITY	Disparity in pixels, expressed at the resolution of the provided map. Positive values = closer. Values <= 0, NaN and -Inf are treated as invalid; +Inf as too close.
DEPTH	Metric depth in sl::InitParameters::coordinate_units. NaN, 0 and negative values are treated as invalid; +Inf as too far; -Inf as too close (same convention as sl::MEASURE::DEPTH).

CUSTOM_CONFIDENCE_CONVENTION

enum CUSTOM_CONFIDENCE_CONVENTION

strong

Lists the value convention of the confidence map optionally ingested with sl::Camera::ingestCustomDepth().

Only consulted when sl::CustomDepthData::confidence is set.

Enumerator
PROBABILITY	Values in [0,1], 1 = confident (typical network output). Default.
ZED	sl::MEASURE::CONFIDENCE convention: values in [0,100], ~0 = reliable, 100 = unreliable.