Efficient City-Sized 3D Reconstruction from Ultra-High Resolution Aerial and Ground Video Imagery

Geo-registering 3D point clouds to 2D maps with scan matching and the Hough Transform

“…The proposed algorithm is compared against state of the art on the Colosseum [1,10], MIT Kendall, and Texas [15] where we attempt to make the tests as fair as possible 2 . In many cases, a preprocessing step was required (to exaggerate edges in the images.)…”

Remote sensing data as a potential source for establishment of the 3D cadastre in Slovenia

“…Nowadays modern technologies of remote sensing, including satellite systems and laser scanning, offer a mass 3D spatial data acquisition of large parts of terrain. A 3D extraction of buildings and traffic routes is being carried out with the use of stereopairs of aerial and satellite high resolution imagery with automatic and semi-automatic processing systems (Long in Zhao, 2005;Gerke in Heipke, 2008;Trinder in Sowmya, 2009;Dornaika in Hammoudi, 2010;Akca et al, 2010;Vasile et al, 2010;Shi et al, 2011;Weng, 2012) or with the use of airborne laser scanning data (Pfeifer et al, 2007;Kada in McKinley, 2009;Pu in Vosselman, 2009;Elberink in Vosselman, 2009;Chen et al, 2009;Tiwari et al, 2009;Wang in Sohn, 2011;Elberink in Vosselman, 2011) that offers a more accurate height determination in comparison to a stereophotogrammetric method (Vosselman in Maas, 2010). On the basis of data from laser scanning, regarded in this paper, this technology proved to be very precise in showing detail on buildings, traffic routes and other objects (Jazayeri et al, 2014), however, it is not appropriate for the determination of property boundary alignment without additional signalizing, except in the case of high point density when it is possible to plot a digital model of terrain of a cell size of no more than 10 cm.…”

“…Danes omogočajo sodobne tehnologije daljinskega zaznavanja, vključujoč satelitske sisteme in lasersko skeniranje, učinkovit zajem prostorskih podatkov v treh razsežnostih na velikih površinah. 3D-zajem stavb in prometnic se izvaja iz stereoparov letalskih in visokoločljivih satelitskih posnetkov s polavtomatskimi ali avtomatskimi postopki (Long in Zhao, 2005;Gerke in Heipke, 2008;Trinder in Sowmya, 2009;Dornaika in Hammoudi, 2010;Akca et al, 2010;Vasile et al, 2010;Shi et al, 2011;Weng, 2012) ali iz podatkov aerolaserskega skeniranja (Pfeifer et al, 2007;Kada in McKinley, 2009;Pu in Vosselman, 2009;Elberink in Vosselman, 2009;Chen et al, 2009;Tiwari et al, 2009;Wang in Sohn, 2011;Elberink in Vosselman, 2011), ki omogoča večjo točnost določitve višin kot stereofotogrametrična metoda (Vosselman in Maas, 2010). Na temelju podatkov laserskega skeniranja, ki ga obravnavamo v tem prispevku, lahko dobro opišemo podrobnosti na stavbah, prometnicah in drugih objektih, vendar ta tehnologija brez dodatne signalizacije ni primerna za določitev poteka posestnih oziroma lastniških meja (Jazayeri et al, 2014), razen pri veliki gostoti točk, ko lahko izdelamo digitalni model reliefa z velikostjo celice, manjše od 10 centimetrov.…”

Mobile Robot Self-Localization Using Omnidirectional Vision with Feature Matching from Real and Virtual Spaces

“…The 3D registration is carried out on two sets of sensor data obtained from sonar, a laser rangefinder or an RGB-D camera. Due to its high computational cost and insufficient accuracy, a number of works have been proposed for further improvements [14,15]. In addition, Salvi et al compared the performance of various registration techniques, including the 3D data obtained from multiple viewpoints [16].…”