Pest monitoring forms the basis of good biological control. Knowledge of the ratio of harmful and beneficial insects is indispensable in determining the necessary control strategy. In the EU PeMaTo-EuroPep project (Pest Management tool for tomato and pepper in Europe), we are developing advanced camera systems and image analysis algorithms for detecting insect pests.
We analysed tomato leaves with spider mite damage with a hyperspectral camera system to investigate which spectral bands contain the most information for distinguishing between healthy and damaged leaves. The results showed that promising classification results can be achieved even when standard colour images are used.
For the upcoming season, a semi-automatic pipe rail trolley has been developed for image and data acquisition in the greenhouse. A deep learning image analysis system that automatically identifies and counts whitefly and bugs on images of yellow sticky traps is also currently under development. The preliminary results are encouraging.