Recent Advancements in Cars AccidentsDetectionby Using Artificial Intelligence Techniques

According to statistics issued by the World Health Organization, car accidents kill more than one million people annually, in addition to property losses. Most car accident deaths occur due to delayed first aid. Therefore, the problem of car accidents has attracted the attention of researchers and scientists. Most of this research aims to design mechanisms for the early detection of car accidents and to inform emergency units in a smooth and fast way to provide rapid assistance to the victims. Many types of research have emerged to detect car accidents, relying on artificial intelligence, deep learning, and various related techniques, such as object detection and object tracking, as well as methods relied on extracting information from sensors such as a speed sensor, acceleration sensor, gas pedal sensor, etc., and then processing this information using artificial intelligence algorithms and techniques.In this research, we propose a simple model based on the training of the YOLOv5 model and the intersection between bounding boxes to detect car accidents. the model that we suggested achieves a 73% accuracy rate with a false alarm ratio of 0.16 through the practical experiments. Our system is designed after a deep review of existing methods, we summarize the most important attempts to detect car accidents and compare the accuracy of each method, its advantages, and drawbacks