Improved Segmentation Techniques for Toxin Detection in Chili Pepper X ray images

Abstract

India is a largest producer and consumer of dry chili pepper. Chili pepper is directly applied to the food without any processing. Inappropriate cultivation, harvesting and storage system of chili pepper leads to decomposition and formation of fungus and it is the source for the aflatoxin. It severely injures the liver and also stimulates tumors in the lungs and kidney. It causes cancer or irritation in the eyes, skin or respiratory system. X-ray method is considered as the simplest and easiest non-destructive methods of detection of toxin in the chili peppers. The research begins its process by capturing X-ray images of Chili pepper, thereafter pre-processing the X-ray image for noise reduction, segmentation includes sub-process of texture region extraction, texture extraction, binarization and morphological operations and the final process is the classification of contaminated or uncontaminated Chili pepper.The objectives of the research are to eliminate the noise from chili pepper X-ray images by using filtering techniques; to evaluate the effect of optimized noise reduction techniques for segmentation process; to extract textures from chili pepper X-ray images by adopting Gabor filter and Texture filters; to enhance texture segmentation of chili pepper X-ray images by employing modifications in the Range filter; to execute binarization process on the segmented chili pepper X-ray images by applying Modified 2D Otsu and Modified Cellular Particle Swarm Optimization; and to detect and remove toxin contaminated chili pepper using Support Vector Machine. newlineMajor objectives : newlineThe research objectives are: newlinei) To eliminate the noise from chili pepper X-ray images by using filtering techniques, newlineii) To evaluate the effect of optimized noise reduction techniques for segmentation process. newlineiii) To extract textures from chili pepper X-ray images by adopting Texture filters, newlineiv) To execute binarization process on the segmented chili pepper X-ray images newlinev) To classify the Chili pepper as contaminated or uncontaminated using Cla