Rubber trees, or scientifically known as Hevea Brasiliensis from Euphorbiaceous family, are particularly prevalent in Thailand, Malaysia, Indonesia, southern part of India and the southwestern part of Sri Lanka . Today, rubber tree has been the main source of natural rubber for use worldwide. It has high export potentials and is rated as one of the most profitable Agri-industrial ventures .
Asian’s well known Rubber tree boards are (RRIM) (Malaysia), RRII (India), RRIC (Sri Lanka), and BPM (Thailand). For decades, these countries have sustained the clone development process for maintaining a successful breeding program. Rubber tree boards have the main role in recommending and distributing clones to the cultivators. One of the important factors that could affect the properties of raw natural rubber is the clonal origin of the rubber tree . The process of clone inspection and verification is very critical because it is important for cultivators to possess rubber clones that produce a high latex yield and having the desired properties in planting of the rubber tree.
One of the ways of clone identification is based on physical characteristics of leaves. However, this clone verification method requires experts with adequate experience. Therefore, the automation of this process is the subject of new research for clone classification.
Plant classification and recognition methods are being performed on plant components such as flowers, leaves, and barks of leaves . Nevertheless, reproductive organs like flowers are unfavorable for plant classification since they are only available for a short period of time in a particular season. Therefore, leaves are more suitable for plant classification. Leaf based plant classification methods are proposed by unique characteristics of leaves including shape, color, and texture .
Existing plant classification systems are unable to classify rubber tree clones due to the leaves have similar physical appearances. However, rubber tree leaves have other sophisticated features that differentiate clones from each other. Leaf tip, leaf base, the form of the leaf or the margin is some of the attributes that might be subject to different feature extraction method in order to generate a clone classification method. The other unique feature is that three compound leaflets are radiating from the leaf base (palmate). These leaflets exist in three different positions; overlapping, touching, or separated. This paper attempts to...