Fraud detection Use Machine Vision while driving

Abbaszadeh, Alireza (2017) Fraud detection Use Machine Vision while driving. Masters thesis, University of Mohaghegh Ardabili.

Preview

Text (تشخیص تخلفات در حین رانندگی با استفاده از بینایی ماشین) A.Abbaszadeh.pdf Download (1MB) | Preview

Abstract

Several factors take part in road accidents: human errors, road problems and technical problems of automobiles. Among these factors, human errors that include driver’s inattention are the most important. National highway traffic safety administartion consider the driver’s inattention as the main cause for nearly 25% of all accidents that includes speaking, eating, smoking, etc. Researches are provided to design driver’s distraction detection system and giving a warn to driver at the right time. In this paper a control system has been proposed that prevents drivers’ speaking,etc. In this method, in order to detect the use of a mobile phone, the first face and the driver’s ear area are detected after the next reduction using the ANN and SVM algorithm, also, to detect the non-use of a seatbelt with image area and referral of the car to the area where the probability of having a seat belt there is the highest percentage after the next reduction using the ANN and SVM is detected. According to the experiments, this algorithm detects 92% of the time in less than a few second on the first image of the use of mobile phones and seatbelts. In 96.8% of cases, in the second image the cell phone and seatbelt are correctly detected and send a textual or audio message to the monitoring unit. Keywords: Machine Vision; Principal Component Analysis; Artificial Neural Network; Support Vector Machines.

Actions (login required)

View Item