Abstract:
Recent advanc
es in computer technology as well as in signal processing and pattern
recognition, provide the possibility of automatic identification of pests, based on their
audio signals. In this research a sound reinforced piezoelectric sensor along with a
detection c
ircuit based on a database was designed in order to receive audio signals with
intensity lower than human hearing limit (zero dB). The confused flour beetle,
Triboium
confusum
was used in this experiment. The signals received from larvae, adult and the
com
bination of these two stages, in wheat grain bulk at three distances of 10, 20 and 30
cm from acoustic sensors were investigated. In each experimental run the characteristics
of signals frequency including range, time, amplitude and intensity were extracte
d. For all
three distances the sound produced by the larvae had a peak intensity in the frequency
range of 2.4 kHz (for feeding) and the adult insect’s sound had two peaks intensity in the
frequency range of 2 kHz (for feeding)
and 2.3 kHz (for walking). T
he differences between
the frequency characteristics of sound produced at different stages might provide the
possibility of identifying the life stages of the pest, pest distance to the sensor and
approximate location of the pest. Based on these findings,
the sound sensor and the audio
circuit were designed to detect larvae, adult, or both at a distance of 30 cm. Further
investigation is continuing to improve the audio system programming and the related
circuits for more accurate detection of the pest.
Keyw
ords:
Acoustic, Insect, Sensor, Stored product,
Triboium confusum
