Determining speed from video for Use of Force and Traffic Reconstruction

Published on December 15, 2017

Kelly Watt

Its amazing the problems video presents to users who are not familiar to the numerous ways CCTV, dash-cam and body-cam recording devices distort the true data. Most proprietary video players distort or change video when the DVR is compressing it, especially when multiple videos are running through 1 DVR, or when a higher quality is taxing the bit-rate.

Images can be stretched as aspect ratios are not accurate, colors distorted, loss of details/edges, strange artifacts show in the image, and/or data is sometimes removed in this process. Often times image frames are dropped or even removed completely from the playable video players. We reviewed one case showing 4 frames per second in this demonstration, when the original frames were 30 frames per second. Also, frame rates can be inconsistent. In one case we showed the original recording with a frame rate starting at 1 frame per second, then jumping to 10 frames per second and showing varying rates. When viewing a use of force scenario it may appear that an officer lunges at the suspect in custody and slammed him against a wall, but after reviewing the inconsistency of the frame rates the story is much different.

Compressed image frames called Predictive frames, or “P” frames, literally ignore changing data. They ignore pixels from the previous frames, only to capture movement. Objects can appear or disappear as a result, distorting what actually occurred causing very poor quality image frames. In order to better understand these video clips or images, we may need to enhance or clarify by removing the “P” frames and matching the “I” frames to better see what had occurred. DVR players interfaces can also cause users headaches with limited functionality or ability to read critical meta data. Investigators screen record video from DVR players or snip screen shots which offer distort results. Traffic Investigators may count image frames with the assumption that frames rates are consistent, presenting results that are not true. Learn more about how you as an officer, investigator, accident reconstructionist or forensic engineers can prevent mistakes during investigations by using the software experts use; iNPUT-ACE.

This demo will show how iNPUT-ACE is able to read proprietary files, extract raw original data in a meaningful way that help you better understand issues that occur when viewing video from DVR’s and proprietary readers. The demo will also show you some good workflows for a better understanding of how speed and frame rates impact our understanding of events in use of force and traffic reconstruction. For a product demonstration, please visit our website or email Kelly Watt