SURF Points Versus SIFT Points in Identification of Medicinal Plants

Pl Chithra,, Professor Department Of Computer Science, University Of Madras,Chennai, India ,S Janes Pushparani,

doi:10.35940/ijeat.a9466.129219

Pl Chithra, , Professor Department Of Computer Science, University Of Madras,Chennai, India + Show 2 more

Open Access

https://doi.org/10.35940/ijeat.a9466.129219

Copy DOI

Abstract

Today, digital image processing is used in diverse fields; this paper attempts to compare the outcome of two commonly used techniques namely Speeded Up Robust Feature (SURF) points and Scale Invariant Feature Transform (SIFT) points in image processing operations. This study focuses on leaf veins for identification of plants. An algorithm sequence has been utilized for the purpose of recognition of leaves. SURF and SIFT extractions are applied to define and distinguish the limited structures of the documented vein image of the leaf separately and Support Vector Machine (SVM) is integrated to classify and identify the correct plant. The results prove that the SURF algorithm is the fastest and an efficient one. The results of the study can be extrapolated to authenticate medicinal plants which is the starting step to standardize herbs and carryout research.

Highlights

Plants are beyond doubt useful for the protection of the environment
The acquired pictograph of foliage vein pattern assists for the input and for the processor to recognize undergoing the procedure of pattern recognition and feature extraction algorithms
The aim of the Support Vector Machine (SVM) classifier is to befit the facts obtained from the Speeded Up Robust Feature (SURF)/Scale Invariant Feature Transform (SIFT) catalogue and to return the finest match based on the pictograph classes of the facts

Summary

INTRODUCTION

Plants are beyond doubt useful for the protection of the environment. It is a complicated and mandatory assignment to recognize the varieties of plants in the land. Identification of leaf contributes to a vivacious role in classification of plants. In contrivance using the leaf and leaf veins, the identification features of the plants are extracted and the mined features are fed-in as basic data for the classifiers to categorize the plants. A catalogue was initiated by means of model imageries for each variety of plants. After the pictograph of leaf is transferred to the personal computer, its indispensable structures are recognized as well as documented by means of image processing methods. This study concentrates by recognizing the title of the herbs by analyzing the foliage vein. Each stint a foliage is recorded, it shall inevitably generate a file christened in the Revised Manuscript Received on December 15, 2019. * Correspondence Author

RELATED WORK

METHODOLOGY

SVM Subroutine

RESULT

Findings

CONCLUSION