Moving Vehicle Detection Using Haar-like, LBP and a Machine Learning Adaboost Algorithm
Salim Al Jabri, Mustapha Saidallah, Abdelbaki El Belrhiti El Alaoui, Abdeslam El Fergougui · 2018
Object detection and classification is one of the core functions of Intelligence Transport Systems (ITS). It is typically based on extracted features and learning algorithms. Different approaches seem to be appropriate. Researchers should compare and evaluate existing approaches to apply the most efficient. In this paper, we propose a moving vehicle-detection vision system. Two solutions are examined in terms of performance and energy-efficient. The first is a classical Adaboost approach based on the Haar-like in feature extraction whereas the second handles a Local Binary Pattern descriptor that will undergo extraction with Adaboost classifier. Comparison results are illustrated based on the GTI vehicle image dataset. The most pertinent is the Haar-like +Adaboost, leading a DR of 90.1% instead of 87.9% for the LBP+Adaboost. However, LBP+Adaboost shows a low energy consumption, which is very important in any embedded systems.