Un primer paso hacia el reconocimiento automático de escarabajos

Beetles belong to the order Coleoptera, and are important organisms for the ecosystems functioning in which they inhabit. Taxonomists classify these organisms to facilitate their identification and study. For this purpose, they use identification keys, composed by dilemmas, which increase in number when classification approaches to the least inclusive category. This work presents a method to automatically identify beetles using as attributes, statistical measures and the Elliptic Fourier Transform. 48 images of 8 different species collected in the Jolalpan region in Puebla state were used. After processing the images and extract the attributes, a classification accuracy of 89.58% was obtained using Random Forest, and 10-folds cross-validation.