Paper
Grading of Bulk Food Grains and Fruits Using Computer Vision
-
Authors:
-
Basavaraj S. Anami; Dayanand G. Savakar
-
Abstract
-
This paper presents a methodology to grade different food grains and fruits image samples using color and textural features. The images of 10 different types of bulk food grains like bengal gram, corn, green gram, groundnut, jowar, peas, red gram, rice, wheat and yellow gram; and five different varieties of fruits like apples, chikko, grapes, mangos and pomegranates in bulk quantity fruits are preprocessed. The color and texture features are extracted and a neural network based classifier is developed to grade the images of different types of produce into four different grades, namely A (Excellent), B (Very good) C (Good) and D (poor). The minimum and maximum recognition and grading accuracies among the food grains are 87% and 95% in respect of wheat and corn respectively and those of fruits are 86% and 91% for chikko and pomegranate respectively by using color and texture features. In this type of approach, more visual features are taken into account and objects are suitably graded. The inherent drawbacks associated with human beings need to be overcome by leveraging the technology. Significant increase in computer processing power and rapid developments in image processing techniques and machine grading are the order of the day.
-
Keywords
-
Color Features; Textural Features; Bulk Food Grain Recognition and Grading; Bulk Fruits Recognition and Grading
-
StartPage
-
1
-
EndPage
-
10
-
Doi
-
10.18005/JAEB0301001