Reconstruction of vehicle speed and acceleration from CCTV data

CCTV cameras and 'dash-cams' are becoming more and more common and regularly record crashes and other incidents. Cameras are used at intersections, worksites and on an array of vehicles. The video data may be accompanied by audio and/or GPS data. When these recorders capture an accident, the data can be used to calculate important attributes of crashes such as vehicle speed and acceleration.

The images from CCTV can be used to reconstruct the position of vehicles and other units involved in the collision. Useful video may be taken from inside a vehicle (e.g. a dash cam) or from outside a vehicle (e.g. a stationary CCTV camera). In order to reconstruct movement from image data, the timestamps associated with each image can be used, along with available waypoint data generated from GPS.

In this case, bus CCTV was used to track its position over time. Using GPS, images from the video, and audio, reliable estimates of speed and acceleration could be derived.

The movement of other vehicles, cyclist or pedestrians outside a vehicle can also be tracked successfully. This usually involves removing image distortion and then tracking movement against identifiable landmarks.

In this case the movement of a cyclists in the seconds before he collided with a support vehicle was captured on video by a spectator. Motion tracking can be used to rectify the images (removing the movement of the camera, blue dots) to extract position of the bicycle providing evidence on speed and braking deceleration prior to the collision.

A range of ancillary data can be used in connection with the images from CCTV cameras, both in their own right, or to provide calibration and validation of the CCTV analysis. The types of techniques Hall Technical have developed for analysing video data will become more widely deployed as more incidents are captured by video systems.