Klasifikasi Batubara Berdasarkan Peringkat Kualitas Menggunakan Faster Region Convolutional Neural Network

Abstract

Coal is a sedimentary rock made up of plant remains that have accumulated through time which can also be used as fuel. Coal has several levels of quality with different functions in each quality. This causes the importance of classifying coal. There are two types of coal classification, namely classification method using technology and chemical analysis classification methods. Coal classification using the chemical analysis method requires a lot of time and cost but has high accuracy results, while the classification method using technology requires little time and cost but the accuracy is not high enough. Therefore, a research was conducted using classification method using technology to find the most appropriate algorithm. This study classifies coal using digital images with coal color parameters combined with the Faster Region Convolutional Neural Network classification algorithm with 3 quality categories; low quality, medium quality and high quality. This study uses 450 data which is split into two categories: training data and testing data. The training data in this study amounted to 360 data and the test data amounted to 90 data. After testing, this study obtained an accuracy of 96,67%, This research can be summarized as being considered quite good in classifying coal with three quality categories; low quality, medium quality, and high quality.