3D mapping with semantic knowledge
read more
Citations
Towards 3D Point cloud based object maps for household environments
Semantic 3D Object Maps for Everyday Manipulation in Human Living Environments
Aligning point cloud views using persistent feature histograms
Simultaneous Localization and Mapping: A Survey of Current Trends in Autonomous Driving
Semantic mapping for mobile robotics tasks
References
A method for registration of 3-D shapes
Least-Squares Fitting of Two 3-D Point Sets
A real-time algorithm for mobile robot mapping with applications to multi-robot and 3D mapping
Learning compact 3D models of indoor and outdoor environments with a mobile robot
6D SLAM with an application in autonomous mine mapping
Related Papers (5)
Frequently Asked Questions (14)
Q2. What are the future works in "3d mapping with semantic knowledge" ?
The aim of future work is to combine the mapping algorithms with mechatronic robotic systems, i. e., building a robot system that can actually go into the third dimension and can cope with the red arena in RoboCup Rescue. Furthermore, the authors plan to include semi-autonomous planning tools for the acquisition of 3D scans in this years software.
Q3. What is the basic idea of the RTS/ScanDrive?
The basic idea of labelling 3D points with semantic information is to use the gradient between neighbouring points to differ between three categories, i.e., floor-, object- and ceiling-points.
Q4. What is the scanner used for this experiment?
The scanner that is used for this experiment is based on a SICK LMS 291 in combination with the RTS/ScanDrive developed at the University of Hannover.
Q5. What is the problem of simultaneous localization and mapping?
Solving the problem of simultaneous localization and mapping (SLAM) for 3D maps turns the localization into a problem with six degrees of freedom.
Q6. What is the value of the value of wi,j?
Given two independently acquired sets of 3D points, M (model set, |M | = Nm) and D (data set, |D| = Nd) which correspond to a single shape, the authors aim to find the transformation consisting of a rotation R and a translation t which minimizes the following cost function:E(R, t) =Nm∑i=1Nd∑j=1wi,j ||mi − (Rdj + t)|| 2 . (1)wi,j is assigned 1 if the i-th point of M describes the same point in space as the j-th point of D. Otherwise wi,j is 0.
Q7. What are the advantages of the RTS/ScanDrive?
As 3D laser scanner for autonomous search and rescue applications needs fast and accurate data acquisition in combination with low power consumption, the RTS/ScanDrive incorporates a number of improvements.
Q8. What is the way to build a 3D scanner?
As there is no commercial 3D laser range finder available that could be used for mobile robots, it is common practice to assemble 3D sensors out of a standard 2D scanner and an additional servo drive [6, 12].
Q9. What is the definition of a 3D point cloud?
A 3D point cloud that is scanned in a yawing scan configuration, can be described as a set of points pi,j = (φi, ri,j , zi,j)T given in a cylindrical coordinate system, with i the index of a vertical raw scan and j the point index within one vertical raw scan counting bottom up.
Q10. What is the way to reduce the equation to a vector?
Eq. (1) can be reduced toE(R, t) ∝ 1NN∑i=1||mi − (Rdi + t)|| 2 , (2)with N = ∑Nmi=1 ∑Nd j=1 wi,j , since the correspondence matrix can be representedby a vector containing the point pairs.
Q11. What is the important improvement in the RTS/ScanDrive?
One mechanical improvement is the ability to turn continuously, which is implemented by using slip rings for power and data connection to the 2D scanner.
Q12. How long does it take to scan a 3D point cloud?
These optimizations lead to scan times as short as 3.2s for a yawing scan with 1.5◦ horizontal and 1◦ vertical resolution (240x181 points).
Q13. Why is the ICP algorithm a problem?
due to the unprecise robot sensors, self localization is erroneous, so the geometric structure of overlapping 3D scans has to be considered for registration.
Q14. What is the optimal translation of the matrices V and U?
Herby the matrices V and U are derived by the singular value decomposition H = UΛVT of a correlation matrix H. This 3 × 3 matrix H is given byH =N∑i=1m′Ti d ′ i = Sxx Sxy Sxz Syx Syy Syz Szx Szy Szz ,with Sxx = ∑N i=1 m ′ ixd ′ ix, Sxy = ∑N i=1 m ′ ixd ′iy, . . . .