Quality Grading of Pea Crop Using Artificial Intelligence

Abstract

In the recent years, computer vision has emerged as a prospective field, related to the recognition of the object by a computer or machine. The most prominent application fields of computer vision are medical image processing, machine vision, military application and optical sorting. In the presented work, the application of computer vision to extract the features of a pea is explored. Feature related to pea are shape, texture, and color. The present work analyse theobject, on the basis of surface areas of the pea, computed from different angle. The quality assigning system based on artificial intelligence is developed. The back propagation neural network is chosen as a quality assigning classifier because of its ability to generate complex decision boundaries. The input to back propagation neural network (BPNN) is range data, consisting of surface areas from different views of object. Surface based analysis technique has advantage that the recognition of object becomes simpler and faster. BPNN uses mean square error as a performance index. The number of hidden layers and number of neurons in a hidden layer are selected on trail basis. The selected network models are simulated with available test data, to evaluate the performance. The result shows the effectiveness of the proposed approach to classify the pea on basis of quality.