|Acquisition Program: ||PMA-290 Multi-mission Maritime Aircraft, ACAT 1|
| ||RESTRICTION ON PERFORMANCE BY FOREIGN NATIONALS: This topic is “ITAR Restricted”. The information and materials provided pursuant to or resulting from this topic are restricted under the International Traffic in Arms Regulations (ITAR), 22 CFR Parts 120-130, which control the export of defense-related material and services, including the export of sensitive technical data. Foreign nationals may perform work under an award resulting from this topic only if they hold the “Permanent Resident Card”, or are designated as “Protected Individuals” as defined by 8 U.S.C. 1324b(a)(3). If a proposal for this topic contains participation by a foreign national who is not in one of the above two categories, the proposal may be rejected.|| Objective: ||Create image processing techniques and automated functions to perform object recognition and imagery matching of disparate imagery sources (ISAR, electro-optic) to enable recognition of maritime targets.
|| Description: ||Disparate imagery sources that rely on different fundamental methods of imagery generation typically reside in military surveillance platforms (SAR, ISAR, electro-optic). Automated and semi-automated methods to extract and match relevant features from these imagery sources are not available. Recognition of maritime targets using these systems typically relies on outdated models of maritime targets to aid identification. The ability to effectively determine appropriate match points from live, real-time imagery will greatly improve surveillance capabilities and greatly improve recognition time. A fundamental problem for computer object recognition is determining correspondence between two sets of image feature from a pair of views of the same scene. This "matching problem" has proved to be very difficult to solve due to the fact that common physical phenomena, changes in illumination, perspective, distortion, etc, can have significant impact on the appearance of a scene and any object in it. This problem is compounded for imagery from very disparate sources such as ISAR and electro-optics.
Additionally, while there has been significant active research into content based image retrieval (CBIR), the ability to search image databases using the content of the imagery instead of textual descriptions of the images in the database, has been focused on imagery from non-disparate sources. CBIR functionality is required by military platforms that need to classify and identify military targets using on-board sensors via autonomous or semi-autonomous means using search functions that do not require extensive image metadata to search. Adding in the additional complexity of disparate imagery sets makes this a research problem that can yield significant military utility with commercial viability.
|| ||PHASE I: Design a methodology that will automatically determine image match points, and identify high interest objects from disparate sources of maritime target imagery to populate a image database that can be queried. Include imagery from ISAR, SAR, electro-optics (EO) sensors, photos and intelligence as sources for features. Determine methodology to query this database using CBIR techniques. Demonstrate the feasibility of improving recognition time for high interest objects.
|| ||PHASE II: Develop a prototype, standard, open architecture database and automated software to extract objects of interest and populate an image database from a wide set of targets. Develop concept of operations for military operators to use real time data to query the image database to identify targets of interest. Demonstrate operator recognition enhancement. Whenever possible, evaluate the performance using sponsor provided data sets.
|| ||PHASE III: Working with the original equipment manufacturer (OEM), transition technology into the Multi-Mission Maritime Aircraft (MMA) via low rate initial production (LRIP) insertion or serial development.
PRIVATE SECTOR COMMERCIAL POTENTIAL/|| ||DUAL-USE APPLICATIONS: The general methods developed could be applicable to a wide range of feature classification needs ranging from those of homeland security to the medical field. Recognition time improvements for detecting objects in disparate imagery using CBIR techniques should be of significant value for homeland security applications in particular.
|| References: ||1. Zhang, S. and Bhanu, B.; “Automatic Model Construction for Object Recognition Using Inverse Synthetic Aperture Radar Images,” Proceedings of the 13th International Conference on Pattern Recognition, 1996.
2. Feng Liu and Jiadong Xu, “Research on Radar Targets Recognition by Extracting 3-D Characteristic from ISAR Images,” Intelligent Information Processing, 2004.
3. Manolis I.A. Lourakis, Spyros T. Halkidis and Stelios C. Orphanoudakis, "Matching Disparate Views of Planar Surfaces using Projective Invariants", British Machine Vision Conference, 1998.
4. Ritendra Datta Jia Li James Z. Wang, "Content-Based Image Retrieval - Approaches and Trendsof the New Age", Association for Computing Machinery, 2005
|Keywords: ||ISAR; Database; Ship Classification; Intelligence; Image Processing; Surveillance|