Inhaltszusammenfassung

We present the sketch of an automatic alternative method for identi-fying and quantifying airborne pollen grains under the framework of a project called Personalized Pollen Profiling and Geospatial Mapping. The first stage of the detection system locates and segments potential pollen particles while re-ducing the dataset size for convenient handling and storing. In the second stage, the proposed classification scheme successively tests different pollen character-istics until the taxon is uniq...We present the sketch of an automatic alternative method for identi-fying and quantifying airborne pollen grains under the framework of a project called Personalized Pollen Profiling and Geospatial Mapping. The first stage of the detection system locates and segments potential pollen particles while re-ducing the dataset size for convenient handling and storing. In the second stage, the proposed classification scheme successively tests different pollen character-istics until the taxon is uniquely identified. Morphological and optical proper-ties of the pollen and image local features are selected based on the study of strengths and weaknesses of the state-of-the-art methods. This scheme allows the system to compute suitable feature vectors to discriminate similar taxa only when necessary, reducing the computational cost. https://hal.archives-ouvertes.fr/hal-00824014/PDF/Lozano_et_al_DGPF_2012_Sketch_of_an_automatic_image_based_pollen_detection_system.pdf» weiterlesen» einklappen

Autoren

Klassifikation

DFG Fachgebiet:
Informatik

DDC Sachgruppe:
Informatik