Car Detection Using Histogram of Oriented Gradients (HOG) Features on Autonomous Vehicle

Autonomous vehicles are becoming a significant innovation trend in the global automotive industry along with the development of automotive technology. This autonomous vehicle concept promises revolutionary possibilities in transportation in the future. In autonomous vehicles, car detection is very necessary to support the car navigation system using the computer vision method. This research aims to propose car detection using a Histogram of Oriented Gradients (HOG) and a Vector Support Machine (SVM). The HOG feature extraction will extract the car image objects, thereby enabling autonomous vehicle systems to identify cars with high accuracy. On the other hand, the SVM classifier is used for the car detection process and separates car objects from other objects besides cars from the HOG features. The detection result showed that car detection accuracy was 83%. This indicates that the car detection in this research is successful and can be implemented on autonomous vehicle navigation systems.