Computer Vision based Model for Fruit Sorting using K-Nearest Neighbour clas...IJEEE
Food grading and estimation has been observed as a key aspect in the field of food and agriculture. Increasing awareness towards quality of food has opened new opportunities of research in this area. Fruit grading and classification is also an important procedure to increase the quality evaluation in fruits grading which affects the export market. Computer vision plays an important role in automation of fruit classification. Total six varities of fruits and vegetable, i.e. red delicious apples, golden apples, green apples, oranges, bananas and carrots are analyzed. The system uses two image databases, one image database for training on the system and other for implementation of query images. In the packaging industry, color and morphological features are the most important feature for classification of fruits. After preprocessing, segmentation is done to extract the region of interest. In this paper, k mean clustering method is used for segmentation to extract region of interest from background. Color features are extracted from the RGB image and HSI image. Morphological features are calculated from RGB segmented image. In this paper, fruits are classified using the nearest neighbor classifier. Euclidean Distance Metric based k- Nearest Neighbor Classifier is developed for this particular application. The overall accuracy of the system is 100%.