DCL containing components responsible for registration of models of objects (SIFTObjctModels - SOMs), their saving to files, loading models from files etc.
DCL depends on the following DCLs:
- CvCoreTypes - contains basic OpenCV-related types
- PCLCoreTypes - contains basic PCL-related types
- PCL - contains algorithms responsible for processing of point clouds (3D vision)
- CvBasic - contains components responsible for "classical" (2D) computer vision
- CvStereo - required for estimation of depth from stereo pairs
- CameraNUI - required for acquisition from Kinect-like sensors
- ModelViewAcquisition - contains component responsible for extraction of masks
- SIFTObjectModel - contains components responsible for generation of SIFT Object Model
Registration of models of objects from stereo camera images:
-
TexturedStereoLRSequenceToCloudViewer - Reads two pairs of left and right images from files (2 with and 2 without additional projection of texture), computes depth and displays resulting XYZRGB point cloud and stores RGB and XYZ files.
-
RGBXYZWideSequenceSOMCollector - Reads RGB and XYZ files along with image acquired from wide camera in order to generate "fragments of object" - views consisting of pairs of XYZRGB and XYZSIFT clouds.
-
CloudXYZRGBSIFTLUMCollector - Loads views (pairs of XYZRGB and XYZSIFT clouds) and registers them, with the use of LUM-based loop closure.
Registration of models of objects from Kinect images:
-
RGBDSequenceSOMCollector - Reads RGB and Depth images from files and generates on their basis "fragments of object" - views consisting of pairs of XYZRGB and XYZSIFT.
-
CloudXYZRGBSIFTLUMCollector - Loads views (pairs of XYZRGB and XYZSIFT clouds) and registers them, with the use of LUM-based loop closure (the same task as for stereo).
Task for calibration of our multi-camera setup:
-
LeftWideSetupCalibration - Task for calibration of a dual-camera setup (i.e. left and wide cameras) with the use of classical chessboard pattern.
-
LeftWideSetupCalibrationCirclesGrid - Task for calibration of a dual-camera setup (i.e. left and wide cameras) with the use of circles grid pattern.
-
RGBXYZWideSequenceDepthTransform_test - Task for testing the transformation of depth (acquired in the left camera frame, transformet to wide, and finally to lazy suzan reference frame).
Additional tasks for displaying images - kinect RGB-D images:
- KinectRGBDRainbowViewer - Displays images acquired on-the-fly from Kinect, enables writing of images to files.
Additional tasks for displaying images - stereo camera images:
-
StereoLRSequenceViewer - Displays left and right images from stereo pair (loaded from files).
-
StereoLRSequenceToRGBXYZViewer - Loads left and right images loaded from files, estimates depth and displays results. Additionally enables to save resulting images to files.
Additional tasks for displaying images - RGB-XYZ pairs of images:
- RGBXYZSequenceViewer - loads RGB-XYZ pairs of images, displays rainbow (depth) and generated XYZRGB point cloud.
Additional tasks for displaying clouds:
-
CloudXYZRGBSequenceViewer - loads and displays sequence of XYZRGB clouds.
-
CloudXYZRGBSIFTSequenceViewer - loads and displays pairs of XYZRGB and XYZSIFT clouds
-
CloudXYZRGBSIFTCorrespondencesViewer - loads and displays two sequences of pairs of XYZRGB and XYZSIFT clouds, estimates and displays correspondences.