Recognition of Fruit and Grading automatically using Machine Learning
International Journal for Research in Applied Science and Engineering Technology
This paper presents an automatic fruit recognition system for classifying and identifying fruit types. The work exploits the fruit shape and colour, to identify each image feature. The proposed system includes three phases namely: pre-processing, feature extraction, and classification phases. Fruit recognition and grading in automatic way is considered as challenging task due to similarities between various types of fruits and external environmental changes. In this paper, fruit recognition and
... uit recognition and grading based on Deep Convolution Neural Network along with keras is proposed. Due to the limitations of previous approaches as had limited dataset and not considered external environmental changes. The final decision was totally based on a fusion of all regional classification using probability mechanism. The results of carrying out these experiments demonstrate that the proposed approach is capable of automatically recognize the fruit name with a high degree of accuracy. Keywords: DCNN, Keras and Feature extraction I. INTRODUCTION Automatic recognition of fruits includes the domains i.e. computer vision and machine learning. This is a work similar to that of classifying images according to their content or detecting objects in natural images but with a more precise and practical use case in mind. Automated harvesting is an emerging field that utilizes computer vision and machine intelligence in order to gather useful information about the growth and ripeness of fruits and vegetables, and other aspects of farming. Computer vision is also utilized in picking and sorting of fruits and vegetables. In this paper, an efficient and less resource intensive approach for automatic fruit classification is proposed. The lightweight nature of the proposed method makes it suitable for embedded devices and single-board computers like the Raspberry Pi. Furthermore, popular techniques for automated fruit recognition are also compared with the proposed method. Nowadays, process automation plays an important role in industries. Many automatic highly efficient methods are developed to use in producing and checking processes. The topic of digital image processing has found many applications in the field of automation. In computer vision and pattern recognition, shape matching is an important problem of which is defined as the establishment of a similarity measure between shapes and its use for shape comparison. A by-product of recognition task might also be a set of point identical between shapes. Shape matching which is intuitively accurate for humans is a needed job that is not solved yet in its full generality. Its applications include object detection and recognition, image registration, and content based retrieval of images. Fruit recognition and classification systems can be used by many real life applications. Such as a supermarket checkout system where it can be used instead of manual barcodes, and as an educational tool to enhance learning, especially for small children and Down syndrome patients. It can assist the plant scientists, where shape and colour values of the fruit images that have been computed can assist them do further analysis on variation in morphology of fruit shape in order and can help them understand the genetic and molecular mechanisms of the fruits. Also, it can be used as aiding tool for eye weakness people which can aid them in shopping as a mobile application. As Fruits play main role in day to day life, grading of fruits is necessary in evaluating agricultural produce. The present existing technology is also used for fruit quality managing purpose but they are not more effective. There are some disadvantages like less reliability, less efficiency and less accuracy. That's why it is necessary to develop a new technology for fruit classification those consist of high accuracy. Analysing the vision is a general characteristic of our brain. Due to advancement in vision based computing capabilities and as algorithms can understand images and videos, systems can be prepared now which understand what we are looking at and what actions we need to perform. Many machine vision algorithms are available for agricultural applications too. A. Problem Statement In existing traditional system, farmer can't identify the actual price for their fruits as there is lack of automatic efficient fruits grading system. The existing systems also has some problems like it is time consuming manual process, third party involvement in between customer and farmers the sellers sell the item with almost three fold of the original price as farmer don't have proper fruits grading system.