Virus detection in tulips: first step towards robotised roguing
Finding virus-diseased tulips in a field is a difficult job. Even experienced hand-roguers find it difficult, in particular in yellow and white cultivars. Another fact is that these professionals are hard to come by, all the more reason to search for different solutions. Sensitive camera techniques seem to enable automation of roguing. The so-called vision techniques seem to offer good possibilities for the detection of virus symptoms.
Applied Plant Research (PPO) and Plant Research International (PRI) started their efforts upon initiative of tulip grower Piet Apeldoorn. PRI has a lot of experience with the development and application of various light-related measuring techniques. This, together with the field experience of growers and PPO, resulted in an experiment in which hand-roguing, a robot, and an ELISA test were compared.
Analysis showed that the robot can observe most differences between diseased and healthy leaves with visible light. This means that an expensive infrared sensor is not required. The results are even better when the optical system of the robot is also ‘looking’ at the forms of the symptoms. In the ‘Yokohama’ cultivar the tested equipment yields the same results as hand-roguing and the equipment was only slightly inferior in ‘Barcelona’! Observation in ‘Monte Carlo’ was extra difficult as result of hail and mouse damage; the low number of TBV-diseased plants also played a role in the final measurements.
Between 30 and 50% of the plants scored as faulty were the same by hand-roguing and the optical methods. This means that 50 to 70% of the faulty scores by the optical method are different from those by hand-roguing! This needs further investigation. Another unaccounted factor is the possibility that vision techniques (cameras) as well as hand-roguing have also scored different symptoms (e.g., a different infestation of the tulip, other than TBV). Fact is that other virus diseases such as TVX or TRV, which may also cause symptoms in the leaf, have not been tested in ELISA.
Photograph: Roguing in the laboratory: measuring TBV symptoms in tulip leaf by means of spectral image information
The Steering Committee considered these research results so promising that a consortium has been formed. This consortium will start follow-up research. A GPS controlled robot may in the foreseeable future be driving through the field, not only detecting and marking diseased plants, but maybe even directly destroying them.