Since the number of traffic accidents in Korea reached a high of 290,481 in 2000, more than 200,000 traffic accidents have been steadily occurring. Classifying drivers' accident factors in the form of traffic violations, safety driving defaults accoun...
Since the number of traffic accidents in Korea reached a high of 290,481 in 2000, more than 200,000 traffic accidents have been steadily occurring. Classifying drivers' accident factors in the form of traffic violations, safety driving defaults account for more than half of all traffic accidents, accounting for 61.3% of all traffic accidents, followed by the Central Line Penetration Act, 6.3% of signal violations, 5.8%, respectively. These statistics clearly show that most accidents play a major role in the failure of safe driving due to the driver's incorrect driving habits. To solve this problem, this study proposes a technology called "Vision-based intelligent driving habit analysis service". The technology detects driving habit defects detected according to various scenarios through object recognition, lane recognition, distance from objects, and relative speed recognition only with 2D images. Defects in the detection contents can be prevented in advance by providing analysis completed images and feedback to the driver on the web to recognize his or her wrong driving habits.