Die Qualität und Quantität von Kohlgemüse kann durch tierische Schädlinge stark dezimiert werden. Schadinsekten werden in der konventionellen Landwirtschaft bevorzugt mittels synthetisch-chemischer Pflanzenschutzmittel bekämpft. Die Applikation von Insektiziden ist jedoch mit Risiken verbunden. Deren Einsatz kann durch Schwellenwerte im Zusammenhang mit Überwachungsmethoden auf ein unbedingt notwendiges Maß beschränkt werden, wie es im integrierten Pflanzenschutz maßgeblich gefordert wird. Die zurzeit verfügbaren Überwachungs¬methoden für Schädlinge im Kohlgemüseanbau unter Freilandbedingungen sind zeitaufwändig und die Entscheidungshilfemethoden nicht flächenspezifisch. Aus diesem Grund sind sie in der Praxis wenig implementiert. Daher war das Ziel dieser Arbeit die Weiterentwicklung und Erprobung der sensorbasierten Überwachungsmethoden zur präzisen Vorhersage des Gefährdungspotentials durch Schadinsekten im Kohlgemüse auf Basis einer digitalen Erfassung, die durch automatische Dokumentation und Interpretation von schädlingsspezifischen Daten gekennzeichnet ist.
Quality and quantity of cabbage crops can be severely decimated by pests. Pest control in conventional agriculture is commonly based on use of synthetic-chemical plant protection products like insecticides. The application of insecticides may, however, be risky because of residues as well as other possible adverse side effects. Therefore, the application of insecticides must be reduced to the absolutely necessary amount as it is required in integrated plant protection. The application of insecticides can be reduced by applying monitoring methods in combination with economic thresholds for pests in vegetable crops. The currently available monitoring methods for pests in cabbage crops under field conditions are time consuming and the decision support systems are not area-specific and for this reason rarely applied in practice. Therefore, the aim of this research was to elaborate monitoring methods to predict the potential risk of cabbage crop pests on the basis of digital detection. Digital detection was characterized by automatic documentation and transmission of pest-specific data.
During the investigations color camera, digital monochrome gray light sensor system, two 3D sensor systems and bolometer camera system as well as CFBA (Chlorophyll fluorescence imaging analysis) were used for the digital detection of unwinged pest groups (lepidopteran larvae or cabbage aphid (Brevicoryne brassicae L.)) directly on cole crops and for the digital detection of plant damages caused by these pests. The gray image camera was examined in connection with three different extraneous light sources, with emitted wavelength of 625 nm, 780 nm or 940 nm. Furthermore, the suitability of three digital sex pheromone lure trap systems (TrapView, Z-Trap-System, a digital funnel sex pheromone lure trap system of the University of Osnabrück), TCS and two imaging light barrier sensor systems (OEOS and Infrascan 5000) for the detection of winged pests were examined. The evaluation of those 12 optoelectronic sensor systems was carried out based on the results generated with these sensor systems, the data collected in the field using common conventional monitoring methods and trials under controlled conditions. The common conventional monitoring methods were crop scouting, sex pheromone lure traps and felt traps.The pest population varied depending on insect species, year and location. Over 56 % of larvae of lepidopteran pests and over 90 % of cabbage aphids were found in the field on the underside of the leaf (facing away from the camera systems). The digital detection of lepidopteran larvae was possible on the top of cabbage crop leafs with color camera system, digital monocrome camera system with extraneous red light (emitted wavelength 625 nm) illumination or CFBA. Large cabbage aphid colonies (>200 individuals) and tissue damages caused by them were best differentiated by CFBA 2D spectral digital signatures. None of the 2D spectral digital signatures could be used to detect small cabbage aphid colonies (<15 individuals) and tissue damage caused by them. Winged cabbage crop pests could be detected digitally with Z-Trap system, TrapView, TCS (Triangel-Camera-System) or OEOS (opto electronic object scanner) system. Image-based digital system TrapView was suitable for automatical detection of diamondback moths (Plutella xylostella L.). The number of diamondback moths in the conventional sex pheromone lure traps corresponded to the number of diamondback moths in the TrapView. Furthermore, the number of larvae on the plants in the field correlated significantly with the number of diamondback moths in the sex pheromone lure traps. Cabbage moths (Mamestra brassicae L.) and swede midges (Contarinia nasturtii Kieffer) could not be digital recorded by the TrapView system. With biosensor of the Z-Trap system, it was possible to digitale detect diamondback moths and cabbage moths under controlled conditions. With TCS it was possible to detect cabbage root flies (Delia radicum L.) automatically during the oviposition. On the basis of OEOS system, diamondback moths, cabbage moths, small cabbage white butterflies (Pieris rapae L.) and cabbage root flies were detected digitally during the flight under controlled conditions.
Thus, image-based digital sex pheromone lure trap systems like TrapView are recommendable for practice for automatic detection of diamond back moth flight activity. Digital recording of winged insect pests using the Z-Trap sytem, OEOS system and TCS as well as digital recording of plant damages caused by insect pests are possible, but these require further research and can therefore not yet be recommended for practice. The digital recording of pests with digital funnel sex pheromone lure trap of the University of Osnabrück, 3D camera system based on time-of-flight measurements (O3D 201PMD 3D), Infrascan 5000 and bolometer camera system were unsuitable due to the insect positioning or insect properties or sensor system properties.
