Information Technology Reference
In-Depth Information
Identify a Specified Fish Species
by the Co-occurrence Matrix and AdaBoost
Lifeng Zhang, Akira Yamawaki, and Seiichi Serikawa
Abstract.
Today, the problem that invasive alien species threaten the local species
is seriously happening on this planet. This happening will lost the biodiversity of
the world. Therefore, in this paper, we propose an approach to identify and exter-
minate a specialized invasive alien fish species, the black bass. We combined the
boosting method and statical texture analysis method for this destination. AdaBoost
is used for fish detection, and the co-occurrence matrix is used for specified species
identification. We catch the body texture pattern after finding the fish-like creature,
and make a judgement based on several statistical evaluation parameter comes from
co-occurrence matrix. Simulation result shows a reasonable possibility for identify
a black bass from other fish species.
1
Introduction
Recently, a number of foreign creature was pointed as a specialized invasive alien
species in Japan. This is because some inconsiderate pet owner discarded their
breeding creature to the outdoor environment, or some fishing fun released the alien
fish species into the river for fishing. Such a happening could destroy the ecosys-
tem of native species, and is occurring world wide. This cause the creatures that
not belong to the new place threaten the local creatures. When a stranger crea-
ture added into a balanced ecological circle, the condition of food sharing, speed
of breeding would cause the native weak species disappeared fast if they are not
be protected correctly. Currently, catching such an invasive alien species manually
is main approach. It will spend enormous manpower and money once the invasive
alien species founded their own life circle, and the effect is not expectable. To avoid
Search WWH ::
Custom Search