Research
Home  >  Research

Laser-based modeling

Study on CPS SLAM - 3D laser measurement system for large scale architectures using multiple mobile robots

In order to construct three dimensional shape models of large scale architectures by a laser range finder, a number of range images have to be taken from various viewpoints and these images have to be aligned using post-processing procedure such as ICP algorithm. In general, however, before applying ICP algorithm, these range images have to be registered to positions close to correct ones manually beforehand in order to be converged to precise positions. In addition, range images must be overlapped sufficiently each other by taking them from dense positions. We propose a new measurement system for large scale architectures using a group of multiple robots and an on-board laser range finder. Each measurement position is identified by the highly precise positioning technique named Cooperative Positioning System, CPS which utilizes the characteristics of multiple robots system. The proposed system constructs 3D shapes of large scale architectures without any post-processing procedure such as ICP algorithm and dense range measurements.


A group of robots equipped with laser range finder, CPS-VIII 3D models by automatic modeling
photo photo
Movie
photo
Movie
photo
Movie
A group of robots equipped with laser range finder, CPS-V 3D map of large scale architecture
photo photophoto
Demo movie
photo
Movie
photo
Movie
photo
Movie
photo
Movie
photo
Movie
Introduction (19MB) Introduction (14MB)
photo
Movie
photo
Movie
Dazaifu Tenmangu shrine Dazaifu Tenmangu shrine
photo photo
Seibuen Park Kyushu University
photo
Movie
photo
Movie
Papers

Laser measurement system of precise tunnel shape

The three-dimensional shape measurement platform of urban architectures such as a tunnel under construction is developed. This system is composed of a mobile platform equipped with a two-dimensional laser scanner, a three-axis fiber optic gyro, corner cubes, integrated software, and a total station. The total station is placed on reference points in a tunnel and measures the position of the corner cubes on the mobile platfrom. Cross-sectional shapes of a tunnel are measured continuously by the two-dimensional laser scanner on the mobile platform, and converted to the three-dimensional tunnel shape according to the position and orientation measured by the total station and the fiber optic gyro. The integrated software is able to compare the measured shape with the design shape loaded from a three-dimensional CAD system via standard LandXML format, and display the error distribution for the management of construction process through GUI in real-time and on-site.


Laser measurement system Multiple robot system
photo photo
Movie
Real-time tunnel shape measurement Measured 3D tunnel shape
photo
Movie
photo
Movie
Papers

Smoothing and impainting technique using laser reflectivity

We developed two denoising techniques using reflectivity for noisy range images: range image smoothing by trilateral filter and range image impainting by belief propagation. The trilateral filter makes use of reflectivity as well as spatial and intensity information so that geometric features, such as jump and roof edges, are preserved while smoothing. The range image impainting technique based on belief propagation recovers a deteriorated range image using not only the adjacent range values but also the continuity of the reflectance image. We conduct simulations and experiments using synthesized images and actual range images taken by a laser scanner and verify that the proposed techniques suppress noise while preserving jump and roof edges and repair deteriorated range images.


Raw range image Bilateral filter Proposed filter
photo photo photo
Shape recognition by slit laser using reflectivity Reflectance image Recognition of reflectance image
photo photo photo
Papers

6D localization using range and reflectance data

6D global localization technique is developed using range and reflectance data obtained by laser scanner and bag-of-features (BoF) technique. Global localization is a fundamental ability to recognize the accurate global position for a mobile robot in a revisited environment. The map-based global localization gives a precise position by calculating an accurate transformation, but the comparison with large 3D data is quite time consuming. The appearance-based global localization which determines the global position by image retrieval techniques with similar structures is real-time. However, this technique needs external illumination constraint and does not work in the dark extremely. We propose a combination of the map-based global localization and the appearance-based global localization. Instead of camera images used for the appearance-based global localization, we utilize reflectance images which are taken as a byproduct of range sensing by a laser range finder.


Proposed technique Pairs in reflectance images using BoF
photo photo
Papers

Geometrical and photometrical modeling of real objects using laser range finders and digital cameras

Tokyo University, Ikeuchi Lab.


Textured image of Kamakura Great Buddha
photo
Movie
Papers