Automated Fruit Quality Detection Using Image Processing

Abstract

In this paper, an automated fruit quality inspection and sorting system for apples has been proposed. The objective of this system is to replace the manual inspection system. This helps to speed up the process, improve the accuracy and efficiency of the system. In this work, the multiple images of apple are captured by using camera place above the conveyor belt. We have captured multiple images to make the proposed system more accurate which covers the maximum surface area of fruit. Then various image processing algorithms are used to extract the features of fruit such as texture, color, and size. Automated sorting and grading are done based on these features. The defected apple is diverted from conveyor using motor. The proposed embedded system works at higher speed and gives high accuracy in grading with low-cost solution. It will have a good prospect of application in fruit quality detection and grading areas. In the proposed algorithm, we have used k-means algorithm for segmentation and local Binary Pattern technique is used extract feature vectors of the test and training images. Then, the training and testing features are given to neural network algorithm to classify the test image into normal and abnormal class. We have considered the normal and abnormal (rotten apple images) for training purposes. Experimental results show that the proposed system accurately classifies the apple and gives better accuracy.