Object Recognition Interface in Vehicles Using Google ML

In this study, the researcher aims to make a ready-to-use system that will detect and recognize vehicles, pedestrians, and traffic signs in an interface for vehicles to add an additional feature for drivers in a traffic environment. Main objectives of this study include the utilization of Jetson Nano module as the main microcomputer and use Google ML kit’s libraries a software kit for the program using newer models from TensorFlow lite. Based on the prototype tested, the system can detect and recognize the three classes for classification under different conditions. However, there are misclassifications of the objects detected due to factors of resolution of the input videos for testing. The overall accuracy of the Google ML kit’s vehicle recognition is 83%.