Abstract:
Extracting and analyzing biological features is becoming a challenging approach for biologists and IT specialists alike. In botany, feature extraction of plant leaf using computer vision would help in classification and early diagnosis of plant diseases. The approach ranges from a simple species recognition using the features of a simple leaf to a sophisticated array of features in the case of a compound irregular leaf morphology. We propose a system for extracting features from an irregular compound-leaf, with minimum user intervention, then establish clustering into distinct similarity groups. We focused on
analyzing features of a tomato compound leaf then a cluster analysis at the variety (intraspecific) taxonomic level was carried out. Experimental results of the clustering process showed that our methodology can be used to classification in the feature. From the samples that were included in the study, two major clusters; potato-type-leaves and tomato-type-leaves were revealed.