Description:
TECHNOLOGY AREA(S): Sensors, Electronics
OBJECTIVE: Design and develop the algorithms needed to perform onboard Automated Feature Extraction (AFE) and/or aided target recognition (AiTR) on Light Detection and Ranging (LIDAR) data.
DESCRIPTION: The U.S. Army has an interest in automated exploitation algorithms for airborne LIDAR systems. These algorithms would take as input, airborne LIDAR data of a specified geographic area. The input data would contain a high-density point cloud created from numerous passes over the specified area. The area may be partially or completely concealed by foliage typical of northern deciduous summer forest environments. As output, the algorithms would produce compressed data products capable of being transmitted via a 2 Mbps data link. The products can take many forms, examples of such products are: Line-of-Communication delineation, void detection, evidence of man-made features, target detection, or 3D scene generation. Ideally, these algorithms would be equally applicable to data generated by any class of LIDAR system (e.g., linear-mode or Geiger-mode).
PHASE I: The Phase I goal is to demonstrate techniques and concepts that could be used to perform AFE or AiTR on LIDAR data. To support development of these algorithms the contractor must provide, or simulate, their own data. The Phase I proposal must describe the data to be used during this phase and give justification as to why this data is valid. The concepts and techniques to be leveraged in Phase II will be demonstrated to government Subject Matter Experts (SMEs). The demonstration of fully autonomous algorithms and real-time hardware is not required during this phase. However, the concepts to be developed in Phase II must be proven.
PHASE II: The Phase II goal is to develop autonomous algorithms for feature extraction and AiTR. Autonomous algorithms refer to the autonomous nature in which the sensor data is ingested and a data product created with no intervention and aiding by a user. The results of these algorithms can/will still require a user to validate the detected target or feature. The algorithms developed in this phase will be based off the approaches demonstrated in Phase I; however, they will be matured to the point of not needing manual interaction. During this phase, the algorithms shall be compared against data sets from a number of scenes and backgrounds to demonstrate performance and robustness to varying scenes and environments. Phase II will conclude with a report describing a detailed description of the algorithms developed, algorithm performance and robustness as well as recommendations for future improvements to the algorithm(s).
PHASE III: The Phase III goal is to take the algorithms from a TRL 5 to a mature state such that they can be transitioned. This includes the system and algorithm improvements described in the Phase II report. These algorithms could then be transitioned to a number of ISR programs. The potential for commercial applications is considerable with such mission areas as search/rescue, mapping, and first responders for situational awareness.
REFERENCES:
1: Peter Cho et al., "Real-Time 3D Ladar Imaging," Lincoln Laboratory Journal, vol. 16, no. 1, 2006, pp. 147–164.
2: Alexandru N. Vasile et al., " Pose-Independent Automatic Target Detection and Recognition Using 3D Laser Radar Imagery," Lincoln Laboratory Journal, vol. 15, no. 1, 2005, pp. 61-78.
KEYWORDS: LIDAR, LADAR, Automated Feature Extraction (AFE), Aided Target Recognition (AiTR), Algorithm Development, Pattern Recognition, ATR, ISR, Event Detection, Onboard Processing
