Separation of Disease Affected Region in Mango Fruits Using Proficient Colorimetric Segmentation

Dr.M. Renuka Devi and Archana Tamizharasan

Due to the change in the ecosystem, there is a sudden rise in the rate of disease infecting the fruit cultivation. Hence diagnosis and detection of diseases in fruits using Image Processing is gaining its importance. Separating the disease affected region from the normal region of an infected fruit generates a pattern and nature to the infected region, which forms the base to classify the disease. There are several techniques available to segment an image but color and cluster based method is important to cluster the pixel based on their density value. In homogeneity is the limitation of cluster based segmentation. This paper proposes a new method of segmentation as Colorimetric Segmentation. This segmentation is tested with mango dataset consisting of images of disease infected mangoes. The accuracy is estimated with Sensitivity, Specificity, Dice Coefficient and Jaccard Similarity as parameters which prove to have better accuracy and maintains the homogeneity. The proposed algorithm efficiency is tested against the existing method and found to have a better result than others.

Volume 11 | 10-Special Issue

Pages: 475-480

DOI: 10.5373/JARDCS/V11SP10/20192834