Intelligent Grading System for Banana Fruit Using Neural Network Arbitration

Abstract

Abstract There is a need for quality production at a very fast rate in food processing industry. Therefore, developing a system that can perform the visual perception of the human operator in making decisions at a very fast rate will be of great advantage. Such machine vision system will reduce human errors such as individual perception differences in determining whether a product is healthy or defective for production. In this research, an intelligent identification system for grading banana fruit has been developed to replace or aid the human operator who may suffer from inconsistent slow decision‐making. This work is divided into three phases. The first phase is the acquisition of the images and preparation of the database required for our experiments. In the second phase, several image processing techniques are employed to extract banana features for use in the last phase; which is the classification phase. Here, a neural network classifier is arbitrated using extracted banana image features in order to classify and grade the fruit. The sufficient classification rates obtained in this work, and the minimal time costs required when compared with previous works indicate that our novel banana grading system can be efficiently used in real life applications in the food processing industry. Practical Applications Our novel grading system has been developed to be used in a fruit (e.g. banana) production factory where quality control and sorting is required. Over the years, human operators had always been employed to grade raw material and product in order to determine if the raw material is suitable for production or marketing. This operation by human workers was considered as very slow when it comes to decision making and there may also be inconsistent in their decision on the product. Thus, we believe that our proposed intelligent grading system can be successfully implemented in practice in a banana production factory, in order to sort out defective or good banana prior to marketing, thus improving the quantity and quality of banana production.2015 Wiley Periodicals, Inc