Automating quantity surveying in road construction using UAV Videogrammetry

Automating quantity surveying in road construction using UAV Videogrammetry

Wortel, T.P.

Khoshelham, K. (mentor)
Gorte, B.H.G. (mentor)

Aerospace Engineering

Optical and Laser Remote Sensing

2009-06-19

Due to changing contract forms in road construction, contractors feel the need for accurate and fast quantity measurement methods. UAV Photogrammetry is a potential answer for contractors. Photogrammetric measurement processing should be automated in order to reach its full potential in speed and obtainability. Automatic extraction of reliable point correspondences between stereoscopic images is still difficult. This research presents a solution to obtain reliable point correspondences between multiple images. The approach is to use Scale Invariant Feature Transform (SIFT) to automatically extract point matches. These point matches are not reliable. SIFT Tracking is a new method that tracks SIFT features in a sequence of images with very high overlap. SIFT tracking, in combination with a motion constraint and a unique matching constraint, extracts reliable point matches. SIFT tracking is tested in road construction image sequences with 99% overlap between consecutive images. SIFT extracts approximately 30 reliable point correspondences over 40 images, such that there is 60% overlap between the first and the ast image of the sequence. These point correspondences can be used to compute a reliable digital terrain model of the road construction site. SIFT Tracking is an improved version of SIFT that can be used to extract reliable point correspondences between photogrammetric images. It fills the last gap in automating photogrammetric measurement campaigns.

Scale Invariant Feature Transform
UAV
Videogrammetry
Photogrammetry
SIFT Tracking
DTM
Road Construction

http://resolver.tudelft.nl/uuid:9a2421c2-9c7b-4e4f-945c-1246ffa35bd4

2009-06-25

Student theses

master thesis

(c) 2009 Wortel, T.P.