Optimal Algorithm for Inner & Outer Region Detection of Indian Pulse Seed Varieties for Feature Extraction and Classification Process

Abstract

The main objective of this work is to develop methodology for identifying varieties of pulse seeds using morphological features. This methodology consists of several steps namely image acquisition, image pre-processing, edge detection (inner region) and image segmentation (outer region) using novel algorithms based on hybrid methods of optimized edge detection with region growing segmentation over Otsu’s thresholding and fuzzy c means segmentation. This paper propose a new automatic and novel algorithm for image feature extraction obtaining by closed segmented boundary using optimized edge detection method by combining canny and polynomial based edge detector. The seeds are developed by homogeneous criteria by incorporating the pixels. This novel approach of efficient segmentation provide homogeneous region with accurate and closed boundaries suitably modified to detect and fill the inner holes. Region growing algorithm is used to obtain all the closed edges in the image. The subject of this study was to investigate the possibility of using novel image segmentation as a tool for pre-processing of feature extraction and classification, designed to extract the morphological features for the Indian pulse seeds. The experimental results shows that the proposed method has gained better promotions in segmentation error rate and running time compared with existing segmentation methods and effective for morphological feature extraction and classification of Indian pulse seed varieties.