Photogrammetry is recognized by many in law enforcement as an important measurement tool for accident reconstruction, since it has the significant benefit of allowing “quick clearance” of roadways during crash investigations. Here, two methods of extracting dimensional information from images are summarized. The first is for three-dimensional information extraction from multiple digital images, using the iWitness^TM system, whereas the second involves two-dimensional measurement from a single image via the software system XYRectify. A demonstration project conducted in cooperation with police departments in Atlanta, GA is used to illustrate the two photogrammetric approaches. Both yield fast, accurate and reliable measurement results from images recorded with an inexpensive consumer-grade camera.

iWitness – A versatile photogrammetric tool for 3D measurement
In January, 2008 DeChant Consulting Services – (DCS), in conjunction with an officer experienced in iWitness use from the Florida Highway Patrol (FHP), demonstrated the photogrammetric measurement approach to numerous Georgia police departments. It should be noted that the FHP utilizes iWitness in their statewide operations as an important part of the “Florida Open Roads Policy.”

The Atlanta Police Department (APD) along with the Georgia State Patrol were interested on this occasion in seeing iWitness used in conjunction with images of a large road intersection, recorded with an Olympus C7070 camera from a low-flying helicopter. The impromptu request amounted to ‘grabbing the digital camera’ and sending an APD officer up in a helicopter to image the scene. Noteworthy is the fact that the officer had never imaged from a helicopter with “photogrammetry” in mind. Also, there was no prior information known of the scene before the aerial imaging; the creation of an accurate 3D model was to be from the 20 or so images recorded with the APD officer’s digital camera of 50mm focal length, along with iWitness. The helicopter flew at an altitude ranging from 250 to 300 feet and the images were recorded in about 5 minutes.

The APD officer also measured a few tape distances within the scene and these were later used to provide accurate scale to the photogrammetrically determined 3D point coordinates. The output results from iWitness are illustrated in a 2Mb video.

XYRectify – A software tool for generating rectified images from which 2D measurements can be made
Rectification is the image processing technique whereby an image taken obliquely to a plane is converted to an equivalent ‘normal’ image of that same plane. An easy way to picture the rectification process is to consider the conversion of an oblique aerial image of the ground to a corresponding vertical image. This process, which requires the establishment of four control points of known XY coordinates in the object plane, along with the marking of the four corresponding image points, is carried out using XYRectify. Following rectification, XY coordinates can be measured within the scale homogeneous rectified image.

It is noteworthy that the rectification process involves a plane-to-plane transformation and thus the resulting 2D coordinates are of optimal accuracy when the object is flat, such as a roadway. Departures from flatness will cause a quality fall-off in the 2D measurement result. In the case of the road intersection, the assumption of flatness can be made for practical purposes, since small variations in ground height will not generate errors of significance given the 1-foot accuracy expectations.

XYRectify formally releases
XYRectify is now included as a free utility program for licensed iWitnessPRO users. One of the key features that sets XYRectify apart from many other rectification programs is the fact that it includes the means to correct for lens distortion in situations where the camera has previously been calibrated. This compensation for lens distortion affords increased measurement accuracy in the rectified image.

There are a number of means to establish control point XY coordinates within the scene. In this case the four control points were measured via the iWitness multi-image photogrammetric survey.

The following screen grab from XYRectify shows the original oblique image, on the left, along with the corresponding rectified image from which XY coordinates can be directly digitized. Once the control point marking is complete in XYRectify, the user simply clicks the “Create Image” button and the rectified image is saved to the working folder as a JPEG image. After rectification, it is a simple matter to “trace” the required evidence of a crash scene from the rectified image, be it generated from a scanned analog picture or a digital camera image, and to export both the resulting XY coordinates and, if required, the rectified image into a CAD system for further mapping.

In the highlighted example, the rectified image (above right) was imported into the Crash Zone CAD diagramming program. Here, any desired rescaling can be performed, though this is unnecessary when good quality control point coordinates are available. The image below illustrates how well an accurate 2D map of features within the intersection overlays the rectified image, thus showing the geometric integrity of the XYRectify processing. The actual 2D diagramming is illustrated below the image. In the case where the scene is actually a multi-car accident site, it is a straightforward process to map critical evidence via both XYRectify for 2D detail and iWitness for more comprehensive and accurate 3D measurements.

Summary
iWitness / iWitnessPRO and XYRectify are low-cost, accurate and easy to use image-based measuring tools for accident reconstruction and forensic measurement. Two methods for mapping from digital images recorded from low-level aerial photography have been summarized here. iWitness (for multi-image convergent triangulation) is more accurate (in this example a few tenths of a foot – when using a consumer grade digital camera.) On the other hand, in many applications XYRectify (single-image rectification) can be used if the required measurements are all on a single plane and the measurement tolerances are not as stringent for the necessary diagramming requirements.

Back to top