Photogrammetry
Written by Cpl. David A. Templeton, Jr.   

MOST LAW-ENFORCEMENT agencies utilize the principles of traffic-crash reconstruction for analyzing the dynamics of collision events. Often, the police officer’s reconstruction work is used in trial and other litigation.

Police have traditionally employed a range of dimensional measurement techniques for traffic-crash reconstruction. These include the roll wheel, steel tape, and—more recently—the electronic distance-measuring total station. The total station has resulted in more comprehensive scene measurements and better diagrams than other methods employed in the past.

Delays in getting the total station to the scene, however, can negate its effectiveness. Even with quick response, the time from incident occurrence to completion of either the traditional baseline-tape method or the total-station method often requires more than two hours—and in some cases may take more than four hours. It is clear that no measurement tool is “perfect” in accident reconstruction and that a complete toolkit of devices is best utilized to perform the task.

A two-car crash near a Wal-Mart store blocked a busy intersection. One of the drivers involved was subsequently arrested for driving without a license. The Florida Highway Patrol used close-range photogrammetry techniques to quickly capture important details at the nighttime scene, resulting in an expedited clearing of the scene and—eventually—a conviction of the driver.

Exploring a new approach

In 2002, the Florida Highway Patrol (FHP) and the Florida Department of Transportation (FDOT) agreed on goals of opening Florida highways within 90 minutes of the time the first officer arrived at the crash scene. This new course of incident management has been identified in Florida as the “Open Roads Policy”.

Goals of the Open Roads Policy include improved officer safety during the crash investigation, a reduction in the amount of time the roadway is closed and a subsequent reduction in traffic delays, as well as lowering concerns of secondary crashes.

To support these objectives, FDOT and FHP representatives did extensive research and attended numerous incident management presentations that were offered by other successful state police agencies around the country. They found that many of the agencies had implemented a technology called close-range photogrammetry.

During the FDOT/FHP research in 2002, the team’s consensus was that close-range photogrammetry allows troopers to reduce their on-scene time by photographing the crash and later “mapping” the crash using a computer back in the office.

The Florida Open Roads Policy does not apply to homicide scenes. It applies only to accident scenes where traffic violations might be involved. The FHP has felt the need to quickly process all crash scenes and begin opening highway lanes as soon as possible. Regardless of the severity of the crash, motorists are generally impatient in long backups and expect the roads to be reopened as quickly as possible. This situation is exacerbated by the need for more accurate scene documentation that will improve the integrity of the investigation and any successful prosecutions. As a result, FHP investigators now find that they need to collect more comprehensive and accurate data in a shorter period of time at the scene.

Utilizing a digital camera as a tool for both measuring and mapping an accident scene, the FHP found that photogrammetry is value-added in supporting the Open Roads Policy because it results in less time on the road compared to other measurement methods used in the past.

The photogrammetry approach


The purpose of this diagram is to illustrate the concept of close-range photogrammetry. Using a film or digital camera, one records multiple 2D (two-dimensional) images of the scene from overlapping perspective angles. Using a PC and photogrammetry software, the user marks the common feature points through a process called “referencing”. The photogrammetry software then computes the 3D (three-dimensional) points through a process called “photogrammetric bundle triangulation”. For a graphic example of how this concept worked on an actual vehicular crash in Florida, view the screen captures on Page 22. (The image shown here was provided by DeChant Consulting Services Inc.)

The term photogrammetry is derived from three Greek words: photos which means light, gramma which means to draw, and metron which means to measure. It is the technology of generating 3D (three dimensional) information from 2D (two dimensional) measurements—and today it is usually done with images that are captured in a digital camera. The process entails taking pictures—either film- or digital-camera images—from different overlapping view perspectives. Images showing multiple perspectives are imported into a PC-based photogrammetry software program where the same entities, using two or more images, are marked—or cross-referenced—to each other. The 3D-coordinate information is derived via a process called photogrammetric bundle triangulation, where the 3D points are computed automatically without requiring the aid or intervention of the operator.

The photogrammetric process determines all of the 3D coordinates. The technology also offers the capability of computing lines and non-contact measured polylines. The operator can add point names to the measured entities, scale the project units, assign a specific coordinate system, and mathematically level the coordinate system to gravity. The result of the image-based measurement work is typically exported in the Drawing Interchange Format (DXF) for diagramming use in a CAD program.

As part of the implementation process, the FHP had tried several different low-cost, commercially-sold, close-range photogrammetry software systems. The department ultimately selected the iWitness program.

Actual case study:
An example of how iWitness works

Although the FHP has successfully demonstrated that photogrammetry can be used on scenes that are tens of meters in overall length,


Close-range photogrammetry software can compute hundreds of points from crash-scene photographs by converting 2D (XY) image coordinates into 3D (XYZ) object-space coordinates for accurate measurement. These coordinates can then be used to create 3D models used for diagramming a crash scene.



This illustration is the final result in a CAD view of images captured with a digital camera and processed in the close-range photogrammetry software program. It took a Florida Highway Patrol trooper three hours in the office to process the images in the photogrammetry software and to create diagrams in a CAD program.

the following example illustrates a two-car crash at night that blocked a busy intersection. The incident involved a violation of right-of-way, where one of the drivers involved was arrested for driving with-out a valid driver’s license. This case was later resolved in a conviction and the mapping of the scene was important in the litigation.

Processing this particular crash scene consisted mainly of measuring photogrammetric natural feature points: crosswalks, signage, vehicles, tire markings, and other critical evidence.

Photogrammetric markers were also used in the measurement process. These markers can be used to measure evidence. However, if the officer can photograph discrete points of interest, those features can be 3D-measured by natural-feature point marking. The markers assist in fast and accurate orientation of the cameras, allowing the software back at the office to easily determine the physical location of each camera when each one of the images was captured.

Time required
for crash-scene mapping

It took the FHP trooper 13 minutes to acquire the pictures of the crash scene. The scene was 193-ft. long by 127-ft. wide. It took another five minutes to set the photogrammetric markers in the scene before the imaging.

Following the crash-scene work, it took one trooper three hours of office time to do all of the photogrammetry and the CAD computer work.

All of the night images were taken with a tripod-mounted camera using time exposures on an average of five seconds each, which produced good image exposure. Seventeen images were used in the photogrammetry work on a PC back in the office. A total count of 202 3D points were measured in iWitness to an accuracy of 0.25-in. RMS (root mean square), as verified by scene-check distances and the program’s auditing tools.

The work done by the iWitness software program was then exported into the Crash Zone diagramming software, where vehicles and other symbols were added to complete the crash-scene diagram.

Value-added technology

According to those who have been associated with it, close-range photo-grammetry has proven to be a value-added tool for the Florida Highway Patrol’s overall mapping capabilities.

Although the traditional methods of baseline tape and total stations are still employed within the FHP, more than 200 FHP troopers have been trained in the image-based photogrammetry technology.

It is now used as a fast, accurate, and easy-to-use tool in crash-scene mapping investigations.

About the Author

Cpl. David A. Templeton, Jr. is a 23-year veteran of the Florida Highway Patrol. In May 1989, he was promoted to the rank of corporal. Templeton is a practitioner of total station and photo-grammetry. During his 19 years as an investigator, he has instructed at various training institutions in Florida and around the United States. He is currently a full-time instructor for Florida Highway Patrol’s Traffic Homi-cide Program. He also provides case reviews and technical support to the investigators in the department. Temple-ton can be reached at: This e-mail address is being protected from spam bots, you need JavaScript enabled to view it

For more information
about the software mentioned in this article, you can go to www.iwitnessphoto.com


ORIGINALLY PUBLISHED:
"Photogrammetry," written by Cpl. David A. Templeton, Jr.
November-December 2008 (Volume 6, Number 6)
Evidence Technology Magazine
Buy Back Issue

 
< Prev   Next >






Item of Interest

The language barrier between English-speaking investigators and Spanish-speaking witnesses is a growing problem. (Updated 28 February 2011)

Read more...