Detection of Texture Anomalies on the Ground (DTAG)

Footprints trails left on unimproved roads during illegal border crossings are often useful starting places for Border Patrol Agents to begin their pursuits. Currently searching for footprints involves CBP agents driving over long stretches of road and visually searching for signs of human activities. In this technology and product development effort, Lynntech proposes to develop a COTS mobile-device based system (smart-phones and tablets) which can utilize vehicle-mounted COTS cameras to automatically scan upcoming segments of road for human footprint trails. At the beginning of the effort the TRL level will be 2, and at the end, the system will be TRL 4. Lynntech has already demonstrated the effectiveness of its proposed footprint detection algorithm on simulated road and footprint imagery during Preliminary Results work discussed in the Technical section of this proposal. In the Phase I, Lynntech will collect actual unimproved road imagery with and without footprints to improve the detection algorithm, and will demonstrate use of the algorithm to detect human footprints at real-time frame rates from a moving vehicle. The proposed approach has dual use as an inexpensive mechanism to enable manufactures to perform real-time surface defect inspection on production lines.