Function of the equipment
The function of the equipment is to detect insects which get trapped in the VARL trap, and to prevent them from escaping, as well as transferring data through the GPRS module to the server.
The equipment has its own power supply for the time between placements, but energy is mainly provided by the solar cell power supply unit.
Structure of the equipment
The equipment consists of the following parts:
- Trap roof with compressed air integrated blow unit
- VARL trap
- sensor case within the trap
- Third generation IR2 sensor
- 12 core, shielded cable for power supply and communication
- D-Sub 15 connector
Structure of the trap
The sensor case is closely integrated into the VARL trap, fastened by 2 screw anchors. In order to prevent pheromone contamination, the closure of the sensor case is ensured by stainless steel, bot hat the upper and lower part, and a glass tube in the detecting band. The stainless steel and glass materials ensure the multi-pheromone usage of the sensor.
The glass design reaches under the probe -using the original funnel system-, thus preventing animals from escaping. During detection the blowing out equipment built into the trap roof is activated, which orients the detected animal to the collecting area with a small pressure (0,5 bar), ensuring clear and correct detection. The trap roof is designed not to be an obstacle for the animals, and it also provides air circulation similar to the original VARL trap. So the pheromone spread is the same as in the case of the original VARL traps, thus providing data comparison of data gathered by electronical and normal VARL trap systems.
The integrated blow equipment built in the trap surface receives power supply and control from the probe electronics. Compressed air (marked in blue) gets to the cover plate from an outer container, through a pressure reducing regulator. Compressed air needed for the blowing out gets into the trap area through a LAVAL nozzle, controlled by the magnetic valve built in the trap roof.
Detection is done by the IR 2 sensor, which has 2 precisely positioned infrared transmitters, and 2×8 pieces infrared sensors (directed toward the transmitter). Its effective detection area is 35 mm, which exceeds the size of the opening of the VARL trap (Picture 1.).
Picture 1. Structure of the trap
The control software is composed of three subunits:
- relational database for permanent data storage
- control unit for tracking the operation of the probes and loggers
- web based application for analysis and display
In the INSECTLIFE project we developed VARL trap-type by adding a sensor into it. With sensors continuous online monitoring of population dynamics of these insects is finally possible. The sensors sends the data through a GPRS system into a central database and data can be seen on a webpage. Because of solar panel energy supply the probe can work for months without any further interventions. Knowing the gradation period of the pest species means planned, cost- and time-effective spray application.
History of probe development
VARL traps were originally constructed for flying pest management (http://www.csalomoncsapdak.hu/). The trap works with pheromones and the captured animals does not find the way out from the construction of the trap and die inside. This trap-type can work during intensive gradation of pest because it has high catching capacity. This design is recommended for catching large amounts of insects. It is especially suitable for catching larger moths (Noctuidae, Geometriidae, etc.) In contrast to sticky designs, it does not loose efficiency when applied in dusty areas (mills, stores).
Development of IR sensors (IR-1, IR-2, IR-3) inbuilt into the VARL traps has been ongoing since 2015.
In the first field season we had to transform the CSALOMON trap. The earlier developed traps work with chemicals which insecticides, which are not allowed in our case. That way animals have time to escape from the trap. We tried a double funnel trap. The first prototype had a totally dark colour (black), to protect the sensoring area from light disturbing the IR sensors (Picture 2.). But it was not attractive for the insects. For the next version except the sensoring area the new probe got a transparent cover (Picture 3.). It was more attractive, but not significantly. Because of the environmental noise, and the hectic movement of the animals, we had to develop other prototypes.
Picture 2. The first prototype
Picture 3. The next version
In the second field season we wanted to drive the animals in a desired direction. First we tried this with a help of a ventilator.
I. Trap with ventilator
In 2016 we used ventilators which blow from the upper part of the probe from an open box. Still we used the dark cover of the sensoring area and a transparent sample container (Picture 4.).
Picture 4. Trap with ventilator at the upper part
After filtering of environmental noise, comparing sensored and counted data we obtained an accurate number of the population size. But the ventilator sprayed the scales from the wings of the animals into the trap, which cause contamination of the sensoring area. Sometimes it killed and ground the animals also, which were attractive for other insect, causing non target and irrelevant sensoring as well.
In 2017 we modified the trap by replacing the ventilator from the top of the trap to the bottom (into a box closed by a net) (Picture 5.). The field tests and observations showed that in many occasion the new ventilator-apparatus begin to work without any detectable reason. In that way we got a far more number of signal than the real number of the trapped animals was. We have more explanation for this failure:
- The moth is able to stay stably on the glass tube which covering the sensor, in spite of the vacuum effect of the working ventilator. The moth is able to move on the tube, which means continuous detection of a single, same specimen.
- In case of successful trapping till it dies, the moth is flittering in the upper part of the container, i.e. in front of the sensor. In that case the sensor provide also signals continuously.
Picture 5. Trap with ventilator at the bottom part
In spite of different data filtering technics used we were unable to find correlation between sensored and observed data.
There were control traps besides the test trap:
- black upper part, transparent sample container, sensor;
- white upper part, transparent sample container, without sensor.
Control probes showed that at the time of field tests population of active plum fruit moth was far larger than the caught value with VARL traps.
At the VARL trap equipped with automatic ventilator, less specimen caused far more false detection signals. Beside the too much signals, the probe had some constructional problems as well. The traps was unattractive for the moths, the animals did not fall into the trap and if it happened they easily found their way out of the trap. It could have several reasons:
- The black, dark, bottom of the trap was unattractive for the animals. Even if they fall into the traps searching for the source of pheromone, they are flittering up towards the light (causing also false signals).
- The too long glass tube which covers also the sensory part hindered the falling of the animals, which instead moving down, flew out from the trap (Picture 6.).
Picture 6. Glass tube in the trap, in front of the sensors
- In case of successful vacuum effect, because the small size of sample container, moth easily find the way out of the trap. Other problem is, that because of the too big ventilator box they are able to rest on it and climb out from the trap (Picture 7.).
Picture 7. The closed ventilator box which is located at the bottom of the trap
According to the above mentioned problems, we tested the probes with a different set up. We made the bottom of the trap transparent (because of the light sensitivity of the sensors we had to use black (dark) upper part) (Picture 8.).
Picture 8. VARL trap version with transparent bottom part
Conclusion deriving from the tests of the new version:
- The colour of the upper part (sensor house) and the long glass tube had no effect on the number of animals caught.
- The ventilator apparatus positively affect their escaping capacity. (The moth sit on it and after that they climb out easily).
- The transparent sample container is more effective, because basically animals move towards the light, in this case into the trap, which results more effective trapping.
II. Traps working with compressed air (blowpipe)
The new type of VARL trap was modified: because of the light sensitivity of the sensor, the upper part remained black (dark) and the bottom part were replaced by a transparent longer plastic tube. Because the version equipped with ventilator was proven ineffective, we used another technics: blowpipe with compressed air. With this method we were able to catch still actively the animals. One of the main advantages of this version is that the tank containing compressed air take place outside the probe (does not occupy place and help the animals escape). In case of animal arrives to the trap the sensor activates the blowpipe which blow the animal into the sample container (Picture 9.).
Picture 9. VARL trap working with compressed air
Two control trap were placed besides the trap:
- black upper part, transparent sample container, simple inner trap, sensor;
- white upper part, transparent sample container, without sensor.
The light sensitivity of the sensor remained a problem (i.e. many false signals).
To solve this problem we stuck a black tape at the upper 5 cm of the transparent sample container. This also hindered the escaping success of the animals (they were unable to find the way outside and rested at the dark part of the sample container) (Picture 10.).
Picture 10. Trap with blowpipe and with black tape at the upper part of the sample container
Due to continuous monitoring we were able to record which signal belonged to actual capture of the trap. With learning machine based on these observation we will be able to run a more accurate data filtering. We will be able to separate actual trapping from environmental noise.
Comparing to control traps, versions with blowpipe (compressed air) caught a bit lower amount of animals, but there were no difference in magnitude of the number of specimens observed using of previous versions.
In case of trap equipped with blowpipe, actual number of caught animals could correlated with the manually filtered sensored data. This type of trap seems to be accurate. However, there is a need to decrease the size of the tank of compressed air to make setting and transporting of the traps easier.