Wavelet energy and shape features for plants recognition

Abstract EN

This work presents plants recognition system with rotation invariant based on plant leaf.

Wavelet energy features are extracted for sub-images (blocks) in addition to three of leaf shape features : [area, perimeter, circularity ratio].

(8) species of leaves are used in different size and color, (15) samples for each leaf are used.

Leaves images are rotated at angles : 90˚, 180˚, 270˚(counterclockwise,clockwise).

Euclidean distance is used, the recognition rate was 98.2% with/without rotation.