Seat Belt Detection Based On The Adaboost

Seat belt detection is a significant subject in the intelligent transportation system. It can warn the drivers who are lack of awareness of traffic safety. Furthermore, it can raise a good habit of driving and reduce damage in the traffic accident. In this paper, we propose a method for lap-belt detection based on the Adaboost We present the main content of our works and achievements as follows:1. We set up a complete system for seat belt detection based on the Adaboost. For the image captured by traffic monitor, we search some candidate regions for lap-belt detection through several detect modules, including the window detect module, the driver detect module.2. For the candidate regions, we extract effective features based on the geometrical characteristic of vehicle and the confidence level in window detection and driver detection. Then we use SVM classifier to find the optimal result.3. The paper also elaborates on typical seat belt detection methods. Starting from some basic theories and algorithm assumptions, it compares the advantages and disadvantages of each method. In the experimental session, the images of multiple scenes in different scenarios are tested to achieve the comparative analysis of experiment results. The experiment results prove that the seat belt detection based on adaboost method perform significantly well.