The dataset is produced within the Safelog project and is used for the evaluation of the developed algorithm which estimates the location of humans in the warehouse environments.
The localization is based on the fusion of information from stereo visual odometry and distances to the known absolute poses of detected ground-markers. The ground-markers are part of the warehouse environment since they are used for the localization of robots:
<img src="./marker.png" alt="markers" width="200"/>
The dataset contains:
./calibration/ -> recordings used for obtaining calibration parameters ./recordings/ -> sequences used for the evaluation ./apriltag_map/ -> recordings of the AprilTags in the environment which are used in obtaining the ground truth pose
The dataset contains 4 recordings made in the Swisslog testing facility in Augsburg. The environment sized 12x13m^2 enclosed with a safety fence contains a dozen metal racks, which is a very close approximation of the real warehouse environment. In the recordings a human worker carried the camera setup on the lower back of the safety vest, while performing expected movements during the warehouse tasks execution.
The recordings are in .bag format and contain the following topics:
/imu -> IMU measurements /stereo/image_raw -> Stereo pair concatenated in one image, left on top of the right image /stereo/left/cam_info -> Parameters of the left stereo pair /stereo/right/cam_info -> Parameters of the right stereo pair /stickers_camera/camera_info -> Monocular camera parameters /stickers_camera/image_raw -> Monocular camera image /tagslam_gt -> Ground truth pose
The images of the testing area and the camera setup are provided in the following images:
<img src="./ag_arena.jpg" alt="Testing_facility1" width="600"/> <img src="./camera_setup.png" alt="cam_setup" width="400"/>
The camera setup contains PercepIn's stereo visual-inertial module Ironsides and the monocular camera FLIR Chameleon3 CM3-U3-50S5M-CS. Both the stereo visual-inertial module and the monocular camera are places on the thick aluminum plate which ensures fixed displacement between them.
distortion: [-0.3495738611503549 0.09447840177522801 -0.003548925540635458 -0.0005183361735231511] projection: [434.1452047256915 434.1526731396501 314.41908249476535 231.01828250185793]
distortion: [-0.35218262885972684 0.09598872982061206 -0.003618221373164676 -2.4773535968930765e-05] projection: [421.7022303158664 421.7236264802341 322.3795386612717 234.81809082346203]
distortion: [-0.05647082111108878 0.34366151497043845 -0.006004431130824592 -0.002859418860702002] projection: [3461.392252592753 3461.1128018631225 1248.4970984850447 980.240992663445]
[0.99993609790634 -0.0026611543382656 -0.0109871907890278 -0.06494060606608117 0.00263031826048 0.9999925643581792 -0.0028200450715273 -0.00030291859693304 0.01099461366738 0.0027909650561792 0.9999356624225201 -0.00029129544625766 0.0 0.0 0.0 1.0]
[ 0.9998361166080204 -0.0087977065772520 -0.0158221454032742 0.03631056812394253 -0.0029816559334763 0.7820179845361501 -0.6232487317194482 -0.04976636403128259 0.0178563617256187 0.6231937677969726 0.7818635930422426 -0.09579991998565489 0.0 0.0 0.0 1.0 ]
The calibration parameters for PIRVS Ironsides stereo-inertial sensor are provided in calib_PerceptIn_V1_2162.json.
The ground truth is obtained with the TagSLAM algorithm which recorded AprilTags in the environment. AprilTags are placed around the environment. The ground truth pose is computed in distinct areas due to the discontinued estimate of the TagSLAM algorithm.
The recordings of AprilTags in the testing facility that are used for obtaining the Apriltags' map, which is afterward used for computing the ground truth, are in the _./apriltagmap folder.