A ROS application to estimate the extrinsic parameters (geometric transformations) between a set of sensors, with respect to a global frame using a ball as a calibration target. The calibration process consists of the following stages:
- each sensor must detect the ball and compute the position of its center;
- the ball is placed in motion in front of the sensors allowing them to detect its center along sucessive positions, creating at the same time a point cloud of detected centers for each of the sensors;
- one sensors is chosen as reference and the remainder are calibrated relatively to it.
This package currently supports the following sensors:
- Sick LD-MRS400001
- Sick LMS151
- SwissRanger SR40000
- Point Grey cameras Flea Gigabit Ethernet (tested with the model FL3-GE28S4-C)
- Microsoft Kinect 3D-depth sensor
calibration_gui is the actual extrinsic calibration package, which requires several other ROS packages:
- colormap provides similar functionality as MATLAB colormap;
- csiro-asl-ros-pkg provides a ROS driver node for the SICK LD-MRS400001 laser scanner;
- RCPRG laser drivers provides a ROS driver node for the SICK LMS151 laser scanner;
- lidar_segmentation provides 2D LIDAR segmentation algorithms. Included on LARtk;
- pointgrey_fl3_ge_28s4_c provides a ROS driver node for the Point Grey FL3-GE28S4-C camera;
- swissranger camera-master provides a ROS driver node for the SwissRanger SR4000 device;
Finally, the multisensor_fusion package is not required by the calibration package (calibration_gui), but is included since it performs multisensor data combination using the previously estimated geometric transformations.
- Clone (including submodules);
- Before compiling, install the following system dependencies: FlyCapture 2.x SDK and libmesasr-dev;
- Compile the packages
ATLAS Project developed at LAR-UA
http://atlas.web.ua.pt/
http://lars.mec.ua.pt/