The Intel® RealSense™ SDK has been discontinued. No ongoing support or updates will be available.
As the first release of the Intel® RealSense™ SDK (Windows) in 2016, R1 (aka version 18.104.22.16828) focuses on improvements for the Intel® RealSense™ SR300 camera, introduction of Platform Camera support, and major changes in Scene Perception (SP) and Enhanced Photograph/Videography (EP). It also provides documentation on how to post a 3D scan to Facebook with Sketchfab. (Please note the downloads now say wm6 or r6 before the actual version number).
New DCMs (drivers) for the R200 and SR300 cameras are also posted.
Requirements: The Intel RealSense SDK supports the following Intel RealSense cameras on Windows: F200, SR300, and R200. Validation used Windows* 10 TH2. Processor requirements vary per camera model.
The Platform Camera is an integrated R200 camera using DCM 2.1.x and specific calibration. More information on the hardware will be published as systems become available.
2016 R1 (version 8) contains changes for SR300 Hand and Cursor modes including improved ranges and motion speeds.
|Full Hand Mode||Cursor Mode|
|Range||+ 15%||up to 90 cm|
|Motion||+ 25%||up to 2 m/sec|
|Calibration||<0.75 seconds||None Required|
|New Features||Click and other gestures|
|Power||1/2 of full hand|
The Hand Cursor interface has also been redesigned to use PXC[M]HandCursorModule, PXC[M]CursorConfiguration and PXC[M]CursorData, independent of the hand tracking interfaces. See https://software.intel.com/sites/landingpage/realsense/camera-sdk/v1.1/documentation/html/doc_hand_cursor_module.html
Both Cursor & Full Hand modes require depth/IR streams with 640x480x60fps. Blob can run on any stream resolution and its default is 640x480x30fps.
NOTE: This 2016 R1 release of the SDK does not support using both FullHand Tracking and Cursor Mode at the same time. This issue will be fixed in the next release.
PXC[M]ScenePerception in R1/v8:
- checks for gravity and inertia sensors: IsGravitySensorSupportEnabled and IsInertialSensorSupportEnabled
- extracts volume projection, vertices, and normal from given pose using GetVolumePreview
- fixed the issue where ExtractPlanes failed/gave error when using any color stream other than VGA/QVGA
AugmentedRealitySP sample now includes all streams supported by SP. The sample shows how to deal with offset in the volume vertices image when color and depth aspect ratios are different.
Scene Perception now supports real-time dense reconstruction, provides access to volume information in real-time, live mesh generation, loading & unloading of scene perception state to & from disk, and provides a utility function to extract planes visible in a camera view. It provides real-time camera pose estimation (tracking & relocalization) in 6DOF using depth, color, inertial sensor support and gravity sensor support in an AR table top scenario with a 2X2X2 meter region of interest on devices equipped with an integrated R200 and platform IMU sensors and specific calibrations .
|New Subclasses||DepthRefocus||DepthRefocus replaced by Init and Apply|
|DepthMask||ComputeMaskFromCoordinate > Init and ComputeFromCoordinate. |
ComputeMaskFromThreshold > Init and ComputeFromThreshold
new MaskParams structure
|MotionEffects||InitMotionEffect > Init|
ApplyMotionEffect > Apply
|PhotoUtils||new API to perform commonFOVPreview|
|New||Depth of field, Common FOV|
6DOF motion effects, two layer and object segmentation,
Layer and object tracking, Crop & Rotate
Depth Enhancement, Color/Depth resize,
|Known Issues||See release notes|
Object Recognition : Now supports
o 34 objects
o Recognition using ROI window
o Localization using selection of 2 object proposals methods
o Object tracking
o Localization and tracking mode
PXC[M]ObjectRecognitionConfiguration interface is extended to configure localization mechanism.
o XML is provided in Plugins
o fixed PXCMImage RefineMask (PXCMPointI32 points, int length, bool isForeground) .
o Enabled IDE IntelliSense giving Parameter info and Quick info about the interfaces. To be refined in future releases.
o Improved 2 expressions: tongue out and puff
o Added 3dscan support
o Enabled Tent (rotation) modes for all face features
o Fixed 3D Detecton flickering issue
o Added more smoothing for landmarks
o Changed default resolution
Face Recognition / PXCFaceConfiguration
To save and read the database file, you must change the sample (https://software.intel.com/en-us/articles/code-sample-facial-recognition-using-intel-realsense-sdk), to comment out the lines shown below:
//recognitionConfig.CreateStorage(DatabaseName, out recognitionDesc);
This is because these APIs are not implemented: RecognitionConfiguration CreateStorage(), SetStorageDesc(), UseStorage() and DeleteStorage()
As in the last release, Speech will NOT work until you load the speech redistributables, which are only provided as a separate install from the Intel RealSense website.
Workaround: Restart the scan from a closer range. The ideal distance from the camera to the front of the object is 30cm for
F200 (or SR300) and 65cm for R200.
The naming convention for the SDK and runtime downloads have changed:
Browser App support is pulled from the SDK and moved to : http://intel-realsense-extension-for-scratch.github.io/
from : http://intel-realsense-extension-for-scratch.github.io/