Low-level Image Segmentation for a Vine Imaging Robot
Simon Flowers
Department of Computer Science and Software Engineering
University of Canterbury
Abstract
Abstract
Image segmentation is an important preprocessing step in most computer vision based appli-
cations, as it can signi cantly reduce future computation in tasks such as object classi cation.
By grouping pixels that are similar with regard to a measure such as colour or position, clas-
si cation can be performed on a per-segment basis, rather than per-pixel. This research
examines several segmentation techniques and evaluates their performance at segmenting the
network structure of vine images. Methods described in the literature are selected for compar-
ison based on their performance at segmenting similar structures. The methods examined are
k-means clustering, mean-shift clustering, normalised cuts segmentation, quadtree segmen-
tation and watershed segmentation. We evaluate each method against ve distinct images,
based on their accuracy and eciency at separating scene components such as vines, posts,
wires and background. Evaluation is performed using a boundary-based comparison method
to compare segmented images against hand generated ground truths. The clustering methods
k-means and mean-shift are found to have the best performance. We propose mean-shift as
the most suitable algorithm, due to its ability to produce a dynamic number of segments. We
provide reasoning behind the relative successes and shortcomings of each method.
