Aggressive Driver Imaging System (ADIS)
By Michael D. Shellem and Joe Schimminger, Instrumentation Development Team, US Army Aberdeen Test Center
While drunk driving or driving under the influence of alcohol is a major concern for law enforcement officials and citizens alike, aggressive driving habits are a problem as well. One way that officials are combating aggressive driving is through the use of systems incorporating laser, video, and PC technology.
The US Army Aberdeen Test Center, as a technology transfer initiative with the Federal Highway Administration, the Maryland State Highway Administration, and the Maryland State Police, developed the Aggressive Driver Imaging System to assist the State Police in identifying aggressive drivers on the Capitol Beltway.
The system uses lasers to determine the range and speed of vehicles on the highway. It then uses a computer system to record video images of the front, side, and rear of a vehicle when the vehicle's measured speed exceeds a predetermined threshold. The information can be quickly assembled into a violation report to be sent to the vehicle owner. A manual override allows the operator to manually trigger the acquisition of video data in order to capture other aggressive driving patterns such as following too closely and erratic lane changes. Figure 1 shows the system installed in a Maryland State Police vehicle.
Theory of Operation
The system uses a laser-type measuring device (LIDAR) to determine the speed and range of an approaching vehicle. The LIDAR and a video camera are aimed out through the rear window of the ADIS platform, a Ford Bronco. The speed is compared to an operator-selected threshold and, if the vehicle speed exceeds the threshold, the system takes a video snapshot of the approaching vehicle and subsequently records a motion video clip of the approaching vehicle. The range and speed are used to estimate the time it will take for the target vehicle to reach the device (AutoSense II) that will trigger the side view and rear view cameras. Figures 2 and 3 give an overview of the system configuration.
When the vehicle reaches the AutoSense II, the system determines whether the vehicle is in the same lane and traveling near the same speed as the vehicle picked up by the LIDAR. If so, the updated speed is used to time the acquisition of images by the side view and rear view cameras. If the vehicle fails to arrive near the predicted time, the system will time out and resume searching for a new target. The AutoSense II also estimates the length of the vehicle. Vehicles over 20 feet in length are categorized as commercial vehicles and the side view camera is used to capture the DOT registration number of these vehicles (Figure 4). Vehicles shorter than this are assumed to be passenger vehicles and the rear view camera is used to record images of the license plate (Figure 5).
Figure 4 (Image has been partly obscured) Figure 5 (Image has been partly obscured)
The operator is able to view live video from any camera by selecting the appropriate switch on the front of the video monitor. Recorded video is displayed on the computer monitor and selected via software.
The system also has the capability of recording traffic statistical information as speed distributions (Histograms) of vehicles in the flow of traffic. These data allow the operator to estimate the average speed of traffic.
Data is saved on a removable disk and subsequently used to generate violation reports that can be analyzed or mailed to the owner of the vehicle. The operator can review any and all data on the computer monitor.
Imaging Developers' Corner
Panasonic NTSC cameras are used for the system, coupled to a Matrox Meteor frame grabber to capture video "snapshots" or "stills" of target vehicles. While resolution and clarity of the images is acceptable, the use of progressive scan cameras in conjunction with a Matrox Corona frame grabber is currently being investigated. A progressive scan camera would enhance image quality by providing full frame, non-interlaced still images. In addition to being able to interface to progressive scan cameras, the Matrox Corona frame grabber has non-destructive graphic overlay capabilities that would allow text and graphics to be superimposed on the live video window.
In addition to using the latest in off-the-shelf, low-cost PC imaging boards, a majority of the software was developed using Microsoft Visual Basic. Matrox's MIL-Lite SDK provided the software functionality for image capture, transfer and display management.
For more information, contact Media Relations.