Ragweed detection based on SURF features

The paper describes a parameter study corresponding to automatic detection of ragweed based on SURF features. The basic idea behind the method is to build a feature database from very simple ragweed samples containing characteristic features of the leaves of the plant, and compare the feature database to features extracted from natural images which contain or lack ragweed. The results of the study clearly show that the approach is promising and has value as a standalone method, or as a potential training basis for a classification expert system.