Growth Behavior of Ragweed (Ambrosia Artemisiifolia L.) on Agricultural Land in Brandenburg (Germany)-Conclusions for Image Analysis in Camera Based Monitoring Strategies

摘要

A requirement to prevent a further spreading of ragweed (Ambrosia artemisiifolia L.), a neophyte to Europe, is an effective monitoring to find single or clustered plants within the natural environment. Noncontact camera sensors attached to agricultural, municipal or aircraft vehicles are an adequate tool to survey large areas of landscape. In result a massive number of images has to be analyzed (if possible in real time) whether ambrosia tissue is present or absent in the respective image. In agricultural fields it would be utile if the imaging is connected with field operations, which had to be done anyway. Subsequently, changing imaging conditions occur. The appearance of the ambrosia and the co-socialized plant species show specific characteristics which can be typically described by metadata. Regarding image analysis various methods exist for object classification which has situation dependent advantages and disadvantages. At the present state of knowledge in the near future an all-purpose classification method under constantly variable environment conditions will not exist. In the paper examples for the appearance of the ambrosia plant in different agricultural fields at different times in the year are presented. According to the situation it follows up a suitable image analysis which classify tissue of the ambrosia. Advantages and disadvantages are discussed. A perspective for a metadata adapted image analysis is shown.