=Paper=
{{Paper
|id=Vol-2030/HAICTA_2017_paper27
|storemode=property
|title=Automated Monitoring of Olive Orchards
|pdfUrl=https://ceur-ws.org/Vol-2030/HAICTA_2017_paper27.pdf
|volume=Vol-2030
|authors=Ilyas Potamitis,Iraklis Rigakis
|dblpUrl=https://dblp.org/rec/conf/haicta/PotamitisR17
}}
==Automated Monitoring of Olive Orchards==
https://ceur-ws.org/Vol-2030/HAICTA_2017_paper27.pdf
Automated monitoring of Olive Orchards
Ilyas Potamitis1, Iraklis Rigakis2
1
Department of Music Technology & Acoustics, Technological Educational Institute of Crete,
Greece, e-mail: potamitis@staff.teicrete.gr
2
Department of Electronics, Technological Educational Institute of Crete, Greece, e-mail:
rigakis@chania.teicrete.gr
Abstract. We present an industry paper on novel insect monitoring appliances
in the field of Information and Communication Technologies in Agriculture,
Food and Environment. We augment typical, low-cost, plastic McPhail-type
traps, with an optoelectronic sensor that identifies the incoming fruit fly from
its wingbeat. The insect counts, environmental parameters, time stamps and
GPS coordinates are transmitted wirelessly from the field, straight to the
remote monitoring agency. We believe that smart traps that report daily the
state of the infestation can, in the very near future, have a profound impact on
the decision making process in crop protection and will be disruptive of
existing manual practices.
Keywords: Bactrocera oleae, Ceratitis capitata, electronic McPhail trap.
1 Introduction
In the context of Integrated Pest Management (IPM), insect pest population
monitoring is crucial [1-3]. The decision of taking action against pests using
chemical or biological measures is based on insect population measurements. These
measurements define the Economic Injury Level; the landmark point in time after
which an economic damage appears. The simplest method to monitor the population
of insects is through the use of insect traps that are commercially available for all
common pests. Insect traps are usually plastic or glass, low-cost containers coming at
different configurations and carrying a pheromone or food attractant. The cost of
applying population monitoring through a network of traps is mainly due to expenses
of manual practices (i.e. wages for placement of traps, scouters that report counts,
zone-managers that pay attention to scouters etc.) As reported in [4], the California
Department of Food and Agriculture operates a network of roughly 63000 attractant-
based traps and in Israel, approximately 2600 traps monitor 20,000 ha of citrus
orchards, both cases against Diptera: Tephritidae. The manual monitoring plan costs
millions of Euros and is a common situation in many countries. We aim at replacing
this manual monitoring procedure with an automated, cost-effective alternative.
Different types of the McPhail trap are commonly used for monitoring and/or
mass trapping of insect populations of fruit flies (Diptera of the Tephritidae family).
The aim of the electronic McPhail trap is to diminish the complicated chain of events
237
related to manual checking of a large network of traps deployed at large spatial
scales. The counts of the captured fruit-flies as well as environmental parameters and
GPS coordinates of the trap are transmitted daily through the mobile network straight
to a central monitoring agency. The central agency can then proceed in the visual
assessment of infestation maps constructed out of interpolating the counts delivered
from the traps. An automated surveillance network is expected to increase credibility
of data, significantly reduce labor costs related to manual scouting, allow timely
gathering of data and reliable situation assessment. The cost of the trap is currently
around 60 € (12/04/2017) for bulk orders and is power sufficient for two months
using rechargeable batteries. In this work, we focus on the industrial characteristics
of the electronic McPhail trap. We elaborate on the design of the trap, its mechanical
components and the assembly and test of the functional prototypes of these traps
currently used for laboratory and field tests.
2 Description of the electronic McPhail Trap
The technical details of the optoelectronic sensor system are described in [6-8].
The housing of the electronics is placed on top of the trap. Fig. 1, illustrates all the
control and processing electronics of the optical sensor. The case is waterproof and
made of white Plexiglas® XT to protect the electronic components against direct
sunlight. In Fig. 2-3 one can see different views of the industrial version.
Fig. 1 CAD design of the electronic McPhail trap.
238
Fig. 2 (LEFT) Embedded electronics. (RIGHT) The electronic McPhail trap.
Fig. 3 Mass production of the electronic McPhail trap has started.
We cover the walls of the McPhail traps in Fig. 3 with a transparent sheet of
plastic on which we apply a transparent thin layer of glue. As a means for
verification data, to validate the automatic counting module, we compare the insects’
stack in the glue and the reported results on the server. We further examine the
recordings that are stored in the SD card to assess the situation. See also Table 1.
Table 1. Open access videos demonstrating the functionality of the electronic trap.
LINKS DESCRIPTION
https://www.youtube.com/watch?v=IdWVaCyHEVI The Electronic McPhail
Trap
https://www.youtube.com/watch?v=7-bKiarPlPs In Lab Experiments
239
3 Inside the trap – Looking at the recordings
Section 3 is directed to the generally knowledgeable and interested reader,
therefore, the description is non-technical. In this section we visualize what the
recordings in the SD card of the trap look-like so that the reader can have a view of
the internal process. The top picture is a typical recording of a B. oleae taken from
the SD card of the electronic trap. We know that is B. oleae positive because we
released a number of them below the trap and this one entered flying-in. In the
second figure we see the spectrum of the wingbeat (i.e. which frequencies constitute
the ‘signature’ of the wingbeat in the frequency domain). The mountain-like structure
is typical of an oscillatory movement. The first peak is the wing-beat frequency
corresponding to the so-called fundamental frequency (f0). One can see that is located
at 200 Hz as expected. This figure is a typical situation of the spectral pattern
originating from a B. oleae. In [8] figure 5, one can verify in another set of
recordings that the wingbeat of B. oleae is a consistent, repeatable and identifiable
pattern. The peaks numbered 1-5 are the so-called harmonics f1-f5 approximately at
integer multiples of f0. One can see that the detection algorithm attributes a high SNR
value to this recording, much higher than 0. The zero threshold is the one under
which a recording is classified as non-B.oleae (i.e. is rejected as being B. oleae).
The third in row figure is a recording of an insect flying in the trap but not B.
oleae. One can again see the structure of a wingbeat (i.e. multiple peaks in the
frequency domain at integer multiples of a fundamental frequency). Note in the 4th in
row figure that the fundamental frequency is around 130 Hz and this is impossible
for B. oleae the beats its wings around 200 Hz. Note that the detection algorithm
attributes <0 SNR to this recording and, therefore, rejects the signal as originating
from B. oleae although it is a perfectly valid wingbeat signal. Last, in the two figures
at the bottom we have the case of an interference. We know that as there is no
wingbeat structure in the signal. The recording cannot be originating out of any
insect, as there is no oscillation. Instead we see a shock-pulse. Note that the
algorithm attributes a large value below zero and confidently rejects the signal as
originating from B. oleae.
Fig. 4 The harmonic detector applied to recordings of the e-trap (Left) A True positive case.
Note the fundamental at 200 Hz. SNR calculation according to the process in Section 3.
(Middle) a non-target signal rejected (SNR<0) for not having the f0 and its associated
harmonics in the spectral area where B. oleae is expected. (Right) a rejected interference
(SNR<0).
240
4 Field results
In this paragraph we present manually verified counting results of all 5 traps
deployed in the field in the island of Crete in Greece. The experiment took place in 1-
12 July 2017, using 5 Entomatic electronic traps (see Table 1). The numbers
correspond to flies-only (manually counted flies against reported number of flies).
During July 2017 in Crete, the temperature was quite high and we did not encounter
B. oleae in the traps. The pheromone traps (B. oleae pheromone dispenser
NOVAGRICA inc.) have been found empty. The device has been switched to higher
frequencies to count flies in general and food-attractant based on gel and
hydrolysable protein. The following are some random files from the SD card. The
symbol ‘T’ denotes Temperature and can be see that the temperature was quite high
(as regards our examples 37.6, 42.4, 43.1 oCelsius). Note also the extremely low
humidity sensor readings denoted with symbol ‘H’. This may explain the fact that we
did not encounter B. oleae in the trap at all, and we therefore switched the algorithm
to detect flies in general. Therefore, results focused on B. oleae only are pending.
F170709_102816_0017_T37.6_H21.3
F170709_120334_0022_T42.4_H15.5
F170709_120154_0021_T43.1_H15.0
Table 2. Summarization of results of all traps deployed in Crete-Greece, in July 2017.
# LOCATION GPS TRUE* REPORTED
1 FANEROMENI Lat:35.0732651, 71 67
Lon:24.8377113
2 CHANIA Lat: 35.50775644 106 126
Lon: 24.0046709
3 ASTRIKAS Lat:35.471774, 135 117
Lon: 23.747486
4 SITIA 1 Lat: 35.194653, 212 202
Lon: 26.110065
5 SITIA 2 Lat: 35.194653, 142 119
Lon: 26.110065
*Manual counting of flies trapped in the glue of the trap.
5 Discussion
We have been observing the traps in the field for several months. Hereinafter, we
summarize our observations regarding their operation in the field:
• The trap does not report false alarm due to sun or other reasons. Although
triggering from non-insect sources occurs at low rates, the recordings produced
by false alarms are successfully rejected by the frequency analysis of its content.
Triggering due to sun appeared only during the hot months of summer.
• The trap has sustained bad weather condition including rain and strong winds
without malfunctioning.
• The detector of the trap discerns the wingbeat of insects and is able to lock on a
specific wingbeat pattern.
241
• There is a very close correlation between insects found trapped inside and
insects counted automatically.
• At its current form, the trap can attract and count flies quite reliably. It can
further focus to B. oleae only if a suitable pheromone attractant is applied or in
orchards where B. oleae is dominant among other flies.
• The device offers the possibility of transmitting the wingbeat snippet to be
classified on a server. In such case, and as reported in [8], there is encouraging
evidence that we could discern B. oleae even with the use of a general purpose
food-bait. This is not investigated yet due to time constraints.
Acknowledgments. We acknowledge Dr. Frank Spiller from the institute for
Microelectronic and Mechatronic Systems in Germany (IMMS) for manufacturing
the housings of the trap and granting permission to use Fig. 1. This research was
funded by EU under grant agreement n° 605073 project ENTOMATIC.
References
1. Oerke, E.C., Dehne, H.W., Schönbeck, F., Weber, A., (1994). Crop Production
and Crop Protection: Estimated Losses in Major Food and Cash Crops. Elsevier
Science. Amsterdam.
2. Flint, M. L., & Van den Bosch, R. (2012). Introduction to integrated pest
management. Springer Science & Business Media.
3. Pedigo, L. P., & Rice, M. E. (2014). Entomology and pest management.
Waveland Press.
4. E. Goldshtein, Y. Cohen, A. Hetzroni, Y. Gazit, D. Timar, L. Rosenfeld, Y.
Grinshpon, A. Hoffman, A. Mizrach, Development of an automatic monitoring
trap for Mediterranean fruit fly (Ceratitis capitata) to optimize control
applications frequency, Computers and Electronics in Agriculture, Volume 139,
15 June 2017, Pages 115-125, ISSN 0168-1699,
https://doi.org/10.1016/j.compag.2017.04.022.
5. Potamitis I, Rigakis I, Fysarakis K (2015) Insect Biometrics: Optoacoustic
Signal Processing and Its Applications to Remote Monitoring of McPhail Type
Traps. PLoS ONE 10(11): e0140474. doi: 10.1371/journal.pone. 0140474.
6. Potamitis I.; Rigakis I. (2015). Novel Noise-Robust Optoacoustic Sensors to
Identify Insects through Wingbeats. IEEE Sensors Journal, 15, no.8, 4621,
4631, Aug. 2015, doi: 10.1109/JSEN.2015.2424924.
7. Potamitis I.; Rigakis I. (2016). Large Aperture Optoelectronic Devices to
Record and Time-stamp Insects Wingbeats, IEEE Sensors Journal, 16, no. 15,
pp. 6053-6061, Aug.1, doi: 10.1109/JSEN.2016.2574762.
8. Potamitis, I.; Rigakis, I.; Tatlas, N.-A. Automated Surveillance of Fruit Flies.
Sensors 2017, 17, 110.
242