DETECTION OF OPTICAL TARGETS IN CLUTTER USING A COMBINATION OF MOTION AND SPECTRAL DISCRIMINANTS

WE PROPOSE TO DEVELOP AND DEMONSTRATE A NEW TECHNIQUE FOR THE DETECTION OF WEAK AIRBORNE TARGETS IN HIGHLY-STRUCTURED BACKGROUND CLUTTER USING PASSIVE ELECTRO-OPTICAL SENSORS. THIS TECHNIQUE, A FEATURE OF WHICH EMPLOYS PROCESSING OF MULTI-SPECTRAL IMAGE SEQUENCES TO EXPLOIT BOTH MOTION AND SPECTRAL DISCRIMINANATS, IS THAT IT NOT ONLY PROVIDES AN INTEGRATED APPROACH TO TARGET DETECTION, BUT ALSO PROVIDES A SYSTEMATIC FRAMEWORK FOR SELECTING THE SPECTRAL BANDS WHICH WILL YIELD MAXIMUM TARGET DETECTABILITY. THE USE OF MULTISPECTRAL INFORMATION SIGNIFICANTLY ENHANCES THE PERFORMANCE OF EXISTING SPATIOTEMPORAL MATCHED-FILTERING (TRACK-BEFORE-DETECT) METHODS. THIS FULLY ADAPTIVE, HIGHLY-ROBUST TECHNIQUE CAN EXPLOIT ANY MEASURABLE DIFFERENCES IN SPECTRAL ENERGY DISTRIBUTION BETWEEN TARGET AND BACKGROUND RADIATION THAT ARE PRESENT AT THE MOMENT OF VIEWING, WHETHER KNOWN OR UNKNOWN. SUCH DIFFERENCES CAN BE CAUSED BY A VERY LARGE NUMBER OF FACTORS, INCLUDING DIFFERENCES IN TEMPERATURE, EMISSIVITY, PROJECTED AREA, REFLECTIVITY, ATMOSPHERIC, ETC. THE PROCESSING TECHNIQUE UTILIZES ALGORITHMS WHICH ARE INHERENTLY INDEPENDENT OF SIGNATURE PHENOMENOLOGY AND WHICH PROVIDE THEORETICALLY OPTIMUM PERFORMANCE. PROOF-OF-CONCEPT STUDIES AND COMPUTER SIMULATIONS SHOW THE POSSIBILITY OF SIGNAL-TO-CLUTTER RATIO ENHANCEMENT IN EXCESS OF 25 DB.