Tracking a depth camera: Parameter exploration for fast ICP

By François Pomerleau, Stéphane Magnenat, Francis Colas, Ming Liu, Roland Yves Siegwart. 2011

119 citations recorded by the source corpus.

The increasing number of ICP variants leads to an explosion of algorithms and parameters. This renders difficult the selection of the appropriate combination for a given application. In this paper, we propose a state-of-the-art, modular, and efficient implementation of an ICP library. We took advantage of the recent availability of fast depth cameras to demonstrate one application example: a 3D pose tracker running at 30 Hz. For this application, we show the modularity of our ICP library by optimizing the use of lean and simple descriptors in order to ease the matching of 3D point clouds. This tracker is then evaluated using datasets recorded along a ground truth of millimeter accuracy. We provide both source code and datasets to the community in order to accelerate further comparisons in this field.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition, Geology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Computer Vision and Pattern Recognition

In Geology

Citations in the corpus

RGB-D datasets using microsoft kinect or similar sensors: a survey (2016, methodology): “The curves under the images are the accelerometer data at 50 Hz of devices attached to the knife, the mixing spoon, the small spoon, the peeler, the glass, the oil bottle, and the pepper dispenser mentioned datasets above, another 26 extra RGB-D datasets for different applications are also added int”
NICP: Dense normal based point cloud registration (2015, methodology): “We performed a comparison of the above mentioned approaches on the benchmark data in [8].”
Kinect and RGBD Images: Challenges and Applications (2012, methodology): “The sensor tracking relative to global model uses a coarse-to-fine Iterative Closest Point (ICP) [25], [26] to calculate the 6 degrees of freedom of the camera.”
Comparing ICP variants on real-world data sets (2013, unknown): “For further details on parameter behaviors, we refer to a previous work (Pomerleau et al. 2011).”
A benchmark for the evaluation of RGB-D SLAM systems (2012, methodology): “To access the color and depth images, we used the openni camera package in ROS which internally wraps PrimeSense’s OpenNI-driver [41].”
SUN RGB-D: A RGB-D scene understanding benchmark suite (2015, unknown): “There are many existing RGB-D datasets available [54] [47] [1] [25] [44] [60] [49] [45] [29] [66] [57] [52] [46] [16] [21] [73] [35] [67] [3] [41] [10] [63] [42] [64] [65] [48] [12] [33] [8] [50] [26] [6].”
Evaluating Egomotion and Structure-from-Motion Approaches Using the TUM RGB-D Benchmark (2012, background): “Pomerleau et al. [39] recorded a dataset with untextured point clouds from a Kinect in a motion capture studio.”
RGBD Datasets: Past, Present and Future (2016, background): “IROS 2011 Paper Kinect Dataset [83] Kinect v1 27 XX ’11”
A Review of Point Cloud Registration Algorithms for Mobile Robotics (2015, methodology): “Having access to a higher frame rate with an optimized ICP solution shows that hand-waved sensor trajectory was trackable with real-time constraints [Pomerleau et al., 2011].”