International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B3, 2012
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia
IMAGE-BASED NAVIGATION OF FOREST HARVESTERS
Marc Schulze
Institute of Photogrammetry and Remote Sensing
Technische Universität Dresden
Dresden, Helmholtzstrae 10, Germany
marc.schulze @tu-dresden.de
www.tu-dresden.de/geo/photo
Commission III/1
KEY WORDS: Photogrammetry, Forestry, Matching, Orientation, Point Cloud, Structure from Motion
ABSTRACT:
The focus of this paper the use of multi-image matching techniques in forestry applications. Background of the study is the problem
of navigating heavy harvesters through skidder trails on their way to harvesting individual trees. Maneuvering these heavy vehicles
over unprotected forest ground leads to irreversible soil compression and degradation effects. Therefore, harvester operators strive to
navigate in a way that exactly the same (already compressed) path is used when they enter a skidder trail for a second time. For this
task, vehicle navigation on a decimeter accuracy level is required. Data of existing techniques, such as GPS, IMU and/or odometry
are error prone, because of difficulties like fluctuating signal strength of satellites caused by dense plant canopy, drift of IMU without
update, and slippery, rough ground for wheel decoding. A camera, as a passive sensor, may avoid these problems, as it is largely
independent to those outer influences.
ZUSAMMENFASSUNG:
Diese Arbeit beschreibt die Verwendung einer Mehrbildauswertung in Forstanwendungen. Der Hintergrund dieser Arbeit ist
die Navigation schwerer Baumerntemaschinen entlang von Rückegassen zu alleinstehenden Báumen. Diese Fahrt über ungeschützten
Waldboden schádigt diesen durch irreparable Bodenverdichtung nachhaltig. Aus diesem Grund muss die Fahrspur eines bereits
gefahrenen Pfades unbedingt eingehalten werden. Dies erfordert eine Fahrzeugnavigation im Dezimeterbereich. Die Daten bestehen-
der Verfahren, wie GPS, IMU oder Odometrie, sind durch Signalverlust, Drift und rutschigen Boden sehr fehleranfillig. Eine Kamera,
als passiver Sensor, konnte diese Probleme vermeiden, da sie grosstenteils unabhängig von äusseren Einflüssen ist.
1 INTRODUCTION 2 RELATED WORK
State-of-the-art forest navigation is based only on GPS and de-
livers an accuracy about 5 to 10m (Hamberger, 2001), caused by
shielding of plant canopy. Vision based approaches might aid
error-prone systems, like GPS and IMU, to navigate in spite of
signal loss or drift effects. Requirements for that kind of naviga-
tion are successively captured images with overlapping regions.
Image position and orientation can be reconstructed by relative
En 2 ! ; 2 orientation based on feature points. Feature detectors such as
existing devices. An image based approach might help to avoid SIFT (Scale Invariant Feature Transform) (Lowe, 2004) or SURF
these problems and increase the accuracy of localization. Cap- (Speeded Up Robust Features) (Bay et al., 2006), are scale and
tured images can be oriented by photogrammetric algorithms and —— jor tion invariant and adequate for tie point detection. Structure
deliver information about motion of used camera. In combina- from Motion (SfM) software packages, like Bundler (Snavely et
tion with GPS, the estimated motion track of harvesters could be al., 2007), take unordered sets of images and produce 3D recon-
transformed into a global coordinate system. Then, all subse- — structions of all camera positions and scene geometry. SfM tools
quent forest machines may follow this more accurately tracked mostly use a sparse bundle adjustment to compute the results and
path to avoid irreversible soil compression beyond the predefined can therefore be used only in retrospect. Another approach re-
skidder trails. : : — ported in (Davison, 2003) offers a real time application for vision
The aim of this survey is to evaluate an image-based navigation based simultaneous localization and mapping (VSLAM). Hence,
for forest harvesters. In a first step, images have been captured — iti, even possible to reconstruct a path with a monocular device.
manually along a nearly straight path within a forest to simulate a
harvester path without vibration and much motion changes. Fast
Modern forest industries need heavy and fast harvest machines
to be economical. Although, vehicles are equipped with mod-
ern navigation systems, forest ground may be compressed and
damaged irreversibly. GPS, IMU and odometry are error-prone,
caused by dense plant canopy, IMU drift and slippery ground.
Therefore, an alternative or an additional system should support
and reliable feature detectors like SIFT and SURF have been used 3 SENSOR AND DATA
to find tie points in overlapping images areas. Further, tie points
have been used to estimate the relative orientation between all For first experiments, manual images have been captured with the
images and to calculate their outer orientation. As a result, the NIKON D700 and a NIKON NIKKOR 20mm wide-angle lens, as
driven track can be visualized as a 3D trajectory, when the images shown in figure 1 and summarized in table 1). Shutter time and
taken along the path are processed in the corresponding order. aperture have been chosen so that motion blur is avoided. The
58
all
er
