The recognition of leaf shapes based on the analytic hierarchy process

Aiming at the plant automatic classification in the precision agriculture, the leaf shape recognition method based on the characteristics of the plant geometric figure was proposed. Several geometric parameters, such as the aspect ratio, lobe number, and invariant moments, were extracted as the features for leaves classification. The analytic hierarchy process was employed to identify leaf shapes. In the first layer of classification, the types of leaf shapes were determined by comparing threshold and the aspect ratio. In the second layer of classification, the broad leaf was classified into single-lobed leaf or multiple-lobed leaf according to lobe number. In the third layer of classification, the minimum distance decision rule was used to identify single-lobed leaf shapes according to their invariant moments. The experiments showed that the analytic hierarchy process can identify leaf shapes effectively and the results of classification were consistent with the human judgment. The leaf shape recognition method has laid the foundation for automatic plant classification. ? 2014 Springer Science+Business Media New York.