HOG AND STATISTICAL FEATURES OFPOTATOES USING MATLAB

Abstract

E-Mail Id: ravindrababu.18@gmail.com 1 Ph.D. Scholar and Professor, Dept. of Processing and Food Engineering, CTAE, Udaipur (India) 2 Assistant Professor, Dept. of Electronics and Communication Engineering, , CTAE, Udaipur (India) 3 Assistant Professor, Dept. of Electrical Enginnering, , CTAE, Udaipur (India) Abstract: Image processing of rotten, crack, good, sprout and skin damage potatoes for extracting Histogram of Oriented Gradients (HOG) is given that in rotten image, rotten part and the areas subjected to initial sprouting are showing good gradients having cell size 18×18, compared to non defect areas of potato. The smooth areas of histogram of oriented gradients (HOG) may be formed due to not using light diffuser while capturing. HOG of cracked and good potato images observed that gradients are rotating in anti clockwise direction. HOG of skin damage shows that gradients at top of the image are weak than other parts of image but initial sprouting images have high gradients. Similar trend observed for sprout images. Contrast of skin damage potato is higher for sprout, rotten, good and crack potato. Rotten and sprout specimen images are equal in correlation values followed by good and skin damage potato. Crack and sprout specimen images contains equal energy values. Rotten, crack and good specimen images contains equal homogeneity values followed by skin damage and sprout potato.