Automatic Vehicle Classification and Counting System Using Inception Model

Abstract

Transportation is very important in life. The width of road is unable to accommodate the total number of vehicles because every year there is a rapid increase in the number of vehicles, causing congestion. In Indonesia, in 2019 the number of motorized vehicles has reached more than 133 million. The process of calculating vehicle volume data which is still done manually has several drawbacks, such as it takes a long time and errors can occur due to human error. In this study, the design of the system used to classify and calculate the number of vehicles automatically utilizes the Deep Learning Convolutional Neural Network with a pre-trained Inception model. The results of this study on the minimum score threshold scenario of 0.4, the highest True Positive (TP) value was 70.75% and the model get 5 FPS during inferencing process.