Abstract

The usage of plants as foodstuff and also medicine is universal. It is important and yet a difficult task to identify the medicinal plants’ species present in the universe through computers. Development of a machine vision system for medicinal plants is a dire necessity as this tacit knowledge dies with the experts in the field. Hence it is helpful to understand and manage these medicinal plants in the interest of pharmaceutical industry and preparation of home medicines. This paper deals with design and development of medicinal plants’ database comprising of 500 species and an efficient text based query interface for retrieval of the desired information. The database consists of scientific and regional names, medicinal values of parts, details of each part such as leaves, flowers and the like, which act as properties or characteristics of plants. The proposed design takes care of access to these medicinal plants through the combination of multiple plants’ attributes. The accuracies of recognition are reported for major and minor properties of the plants.  The results have revealed that leaves and their shapes invariably help in recognition of plants. Hence, Furthermore, retrieval of relevant information through content based approaches is the ultimate target.

Key words: Database, Indian medicinal plants, query processing, image retrieval.