Speed Determination Using Audio Analysis of Dash Camera Video for Vehicle Accident Reconstruction

Video from dash or surveillance cameras is sometimes used in vehicle accident reconstruction to analyze the speeds of vehicles. However, video captured during nighttime, during poor visibility conditions, or of events out of frame may not always visually capture details needed to determine the speed of the vehicle in question. Prior research has determined speed from vehicle acoustic signals, but little research has analyzed the audio portion of dash camera video for use in accident reconstruction and other forensic settings. The purpose of this study was to outline and test the validity of a method for using the audio portion of dash camera video to determine vehicle speed. Extracting the audio portion from the video recording and further processing it with commercially available software can allow the calculation of vehicle speed and acceleration when traveling over roadway surfaces and detection of turn signal activations while driving. By extracting the audio portion from the recorded video and processing in the frequency domain to remove ambient noise, the speed of the test vehicle was calculated when the wheels of the test vehicle traveled over expansion joints, rumble strips, and grooved concrete roadway surfaces. The reliability of the audio data were then tested against the video and speed data collected for a baseline. The proposed method can be useful to accident reconstructionists in that it provides an additional method for their analysis of vehicle speeds.