“Photogrammetry” Is Sharper Than The Eye

The headline was simple – another person exonerated in a case where police claim he had confessed and where two eyewitnesses positively identified him.  But absent from the story was any mention of DNA as the exculpating tool.  Instead, it was reported as follows:

The experts reviewed surveillance footage of the botched heist of Northeast Philadelphia’s William Glatz Jewelers — which ended in a shootout that killed both Glatz and robber Kevin Turner — and concluded the perpetrator was at least four inches shorter than Onyiah. That aligned with the height estimates given by all four eyewitnesses on the scene in 2010.

Melamed, Lawyers say Philly cops coerced a man to confess to murder. He’s cleared after 10 years, https://www.inquirer.com/news/philadelphia-ciu-exoneration-da-larry-krasner-patricia-cummings-obina-onyiah-20210504.html (last visited May 8, 2021).  The tool used was photogrammetry.

This was the second instance known to this author in which measurement using technology showed that a person named as being the individual in a crime scene photo was wrongly identified.  In 2013, Tahmir Craig was released after ten months in jail where he sat charged with a murder he did not commit.  When a photo from a surveillance camera was broadcast to the public, a store owner acquainted with Craig identified him as the individual leaving the murder scene.  After the photo was analyzed by the FBI, “it was determined the shooter’s height was 5’11; Tahmir Craig is 5’5.”  Man wrongfully arrested in Chester murder freed after 10 months (March 20, 2013), https://6abc.com/archive/9035084/ (last visited May 8, 2021).

The process is not new, and in one of the first reported instances of its use it was prosecution evidence:

By analyzing two photographs, FBI Agent Douglas Goodin concluded and testified that the Berkeley robber was between 5’3” tall and 5’6” tall. Quinn is 5’5” tall. To determines the Berkeley robber’s height, Agent Goodin used a process in which a formula is derived by measuring the change in the dimensions of objects in a photograph as they move away from the camera. After testing the formula against objects of known dimensions in the photograph, Goodin was able to make an estimate of the robber’s height.

United States v. Quinn, 18 F.3d 1461, 1464 (9th Cir. 1994).  Indeed, one author traces its origins to Leonardo DaVinci and the study of perspective.  And like DNA evidence, what was first used as a means to support a conviction is now utilized to show actual innocence.

The general reliability of photogrammetry evidence is seemingly well established.  “Federal and state courts have recognized photogrammetry as being a reliable science and have allowed expert testimony on the subject…”  Kittle v. St. Jude’s Co., 2010 Colo. Dist. LEXIS 2037, *3   At the same time, photogrammetry evidence is not always proper.  A recent decision reversed a trial court’s admission of prosecution photogrammetry testimony because the original photograph, shot at night and less than “pristine,” left too many variables unknown.  The expert admitted that

there was “incalculable uncertainty” based on factors such as the quality of the image (it was taken at night), the unevenness of the terrain, the body position of the individual, the inability to see his feet, and the head covering the individual was wearing. She expressed concern about this particular examination because the subject was standing a “considerable distance” from the camera, felt obligated to “qualify” her results because the uneven terrain would make a difference in whether the camera lens height was actually at the position of the suspect, and noted that “because the terrain was somewhat uneven it wasn’t able to be completely, accurately collected.”

Matthews v. State, 249 Md. App. 509, 539, 2021 Md. App. LEXIS 158, *38, 246 A.3d 644.  In another instance, such testimony was repudiated because it came from a detective who “made no calculations. Instead, he made one measurement using a tape measure and testified to the defendant’s height based upon what appears to be an extrapolation from that measurement.”  Williams v. State, 300 So. 3d 202, 209, 2020 Fla. App. LEXIS 10193, *16, 45 Fla. L. Weekly D 1690, 2020 WL 3982828.  The Florida court rejected this as inadequate to satisfy the threshold of expert testimony and as non-helpful lay opinion evidence.

Back to the exoneration of Mr. Onyiah, a man 6’3” tall and weighing 195 lbs.  At the time of the crime, the two eyewitnesses described the perpetrator as 5’8” tall.  The police then developed a lead that the doer might be a person named Donte Waters.  Yet this information was discarded when a prison inmate told police that he “saw the video from the robbery on the TV three times and the male in the video was Obina Onyiah.”  Commonwealth v. Onyiah, Joint Stipulations of Fact, 27.  Brought into custody, a confession to the crime was obtained.  For trial, neither the prosecution nor defense counsel consulted with a photogrammetry expert.  Id.

Attempting to conduct measurements a decade later was problematic – the store had closed so shelves, display cases and other objects could not be measured for comparison with the suspect.  But what proved available were the heights of known individuals also in the crime scene photographs.  Three experts, working independently, reached the same conclusion – the male in the store video could not have been Mr. Onyiah and was several inches shorter.

Here is the surveillance photo of the perpetrator, followed by one of the comparison photos used by one of the experts:


What are the lessons?  Several.  A police confession trumped significant height disparities; available technology was not sought out when it could have been – either by police investigators or defense counsel; video capture of crimes offers critical proof but needs to be analyzed scientifically; and yet again, eyewitness testimony has been shown to be stated with confidence and convincing but indisputably wrong.