Segmentation of fruit and vegetables using fuzzy - Corner based region growing algorithm (Fuzzy-COBRA)

Abstract

This paper proposes a Novel algorithm called Fuzzy - Corner Based Region Growing Algorithm for shape analysis and object segmentation in Fruit image processing and specifically for Detection of Fruits and Vegetables in the large farms. The proposed research focus on extracting the corners of the object with varying shapes using the proposed Fuzzy-COBRA. It is used in corner and edges at the initial part and those corners are fuzzified and processed with region growing algorithm on input images. The process starts from an input image imported to system using imaging devices directly or benchmark Kaggle fruit Image Dataset and FIDS30 to detect the loss of information in the edges and corners of the imported edge due to noise and position offset. Proposed algorithm recognizes and removes false noise using surface filters and some acquisition tools to reach the accuracy level in detecting the shape of fruits, and vegetables. It solves the extreme points of the intersecting line of curved points. Proposed a Fuzzy-COBRA positions the corners of the imported image to capture important events and improves the object quality. The experimental results show the improve in efficiency to in segmenting the fruit and vegetables by detecting the corners after fuzzification and achieves higher accuracy in detecting the position of the edges from images of fruits and vegetables especially in organic agriculture during export and import in large scale.