Mapping city environments using a single hand-held digital camera

Mapping city environments using a single hand-held digital camera

Densely populated cities pose a problem for reliable navigation and tracking solutions using GPS. Narrow streets between high-rise buildings blocking GPS signal paths provide limited visibility to satellites and cause multipath effects, resulting in degraded navigation accuracy and reliability. In this context this paper proposes a conceptual framework for a mapping process in an attempt to survey city blocks where the only sensory input is a single low-cost digital camera.
This proof-of-concept measuring process begins with taking images of at least two known geo-referenced control points and involves taking sequential images as the user walks forward around an city block. Each image is linked to previous images of the same scene taken from a previously occupied location.
Bundle adjustments and registration algorithms compute the direction and orientation between sequential images that, in turn, allows the extraction of 3D information of selected target points visible in the images. Computations are carried out within an office environment using a consumer-grade photogrammetric software program.
A controlled field experiment is described to demonstrate the capabilities of the system. This experiment is not intended to be comprehensive, but it gives useful results that can be taken as typical indicators of the accuracy to be expected from the proposed measuring scheme. This was ascertained by statistically evaluating and comparing the coordinates’ values obtained during the test phase and the coordinates of the same target points as computed with what was considered to be more accurate total station surveying instrumentation.

File reproduced in accordance with the copyright policy of the publisher/author.