Haar-like feature based real-time neuro car detection system

Abstract

A real-time neuro car detection system based on the Haar-like feature is presented in this paper. The proposed system relies on an artificial neural network (ANN) to recognize the car object. ANN was trained using the Haar-like features extracted from the negative and positive car image data. The car objects vary with their sizes and trademarks. However, they have common features which can be assumed unique for the car. In this paper, the common features of the various car objects were transformed into the Haar-like features and then used to train ANN. The system was implemented on the embedded PC Raspberry Pi 3 using the camera SJCAM SJ4000. The research results show that the detection accuracy was influenced by many factors. The developed system resulted in the accuracy coefficient of up to 95% and the detection speed of about 700 ms per frame.