Purdue researchers, DOJ-funded, prototype portable 3D imager for footwear and tire impressions

RTI International Forensic Technology Center of Excellence webinar · February 4, 2026

Loading...

AI-Generated Content: All content on this page was generated by AI to highlight key points from the meeting. For complete details and context, we recommend watching the full video. so we can fix them.

Summary

A Purdue University team demonstrated a DOJ-funded prototype for an affordable, portable 3D imaging system to capture footwear and tire impressions; the presenter said software will be free and the target device price is about $10,000, with production possible next year.

Purdue University mechanical engineering professor Dr. Song Zhang demonstrated a Department of Justice–funded prototype for a portable 3D imaging system designed to capture footwear and tire impressions, saying the project’s goals include high-resolution capture, fully automated processing and an affordable end product for field examiners.

The project, Zhang said, was initially funded by the Department of Justice and has moved into a second phase of development. “The objective of this work was, try to develop a high resolution portable 3D imaging technology for footprint and tire track capture,” Zhang said during a webinar hosted by RTI International’s Forensic Technology Center of Excellence. He told attendees the team is aiming for a product price of about $10,000.

Why it matters: investigators and crime-lab examiners currently use casting and 2D photography to document impressions. Zhang and his collaborators, including retired forensic examiner Jim Wolf, argued that digital 3D captures can preserve fine surface details that casting sometimes misses and can avoid casting artifacts that obscure small features.

How it works: Zhang described a structured‑light approach. A projector casts patterned stripes onto an impression, a calibrated camera observes distortion in those stripes, and triangulation reconstructs the surface in x‑y‑z coordinates. The team built software to automate exposure and to choose between single‑shot processing and a high‑dynamic‑range (HDR) workflow when scene contrast requires it.

Prototype performance and limits: Zhang reported the early prototype captures about 25 images per 3D scan and that a raw 3D scan in the team’s internal format is roughly 18 megabytes; exported standard 3D files (for example, OBJ-type formats) can reach about 300 megabytes. He said low‑resolution prototype captures were roughly 137 dots per inch (dpi) and that a high‑resolution configuration (around 360–400 dpi) produced materially clearer small features. “We were concerned that in very fine cohesive soils a dental cast sometimes resolved more detail than the lower‑resolution 3D capture,” Zhang said, adding that the higher‑resolution option restored confidence in capturing small cuts and bumps.

Software, workflow and hardware design: Zhang said the team is designing a plug‑and‑play box containing two cameras and a projector, mounted on a tripod. The software includes a visualization pane that can mirror the 3D geometry (a “digital cast”), export mesh files and save screenshots. He told the audience the software is planned to be free: “the software will be free to anyone to use,” Zhang said.

Field practicality and feedback requests: Zhang emphasized automation to reduce training burdens for field examiners. He asked practitioners for input on preferences such as battery architecture (user‑replaceable vs. built‑in), Wi‑Fi or touchscreen control, weight targets (the prototype is about 5 pounds), and operating temperature requirements. He said hardware calibration would be performed before shipment so end users would not need to recalibrate unless cameras or lenses are changed.

Questions from the audience addressed several operational issues. Zhang recommended a tripod for stability because handholding can introduce blur when tens‑of‑micron detail matters. He said a manual capture can take about 0.5 second with processing on a laptop in roughly 2 seconds for a single capture; HDR capture currently requires about 6–7 seconds. Zhang acknowledged large file sizes for stitched or long impressions and said his team is developing compression and stitching algorithms to create larger scans from multiple high‑resolution tiles.

Availability and next steps: Zhang said his team has funding to pursue product development and anticipates production could begin as soon as early next year if practitioner requirements are finalized. He said the group will ship prototype units to practitioners for testing and will solicit feedback to refine the product.

The session concluded with instructions to download the PDF slide deck, a reminder that the webinar recording will be posted within days, and a request that attendees complete a short survey to inform product decisions.