Abstract: Methods and apparatus for processing data from multiple sources according to various aspects of the present invention include a sensing system for determining target information. The sensing system includes sensors to generate target information and local trackers to receive the target information from the sensors. The local trackers may generate local target data for the target, and a combiner connected to the local trackers may use the local target data to generate global data for the target.

Abstract: Methods and apparatus for processing data from multiple sources according to various aspects of the present invention include a sensing system for determining target information. The sensing system includes sensors to generate target information and local trackers to receive the target information from the sensors. The local trackers may generate local target data for the target, and a combiner connected to the local trackers may use the local target data to generate global data for the target.

Abstract: A generalized Hough transform is used to acquire and track vehicular targets for missile guidance. This is accomplished by recognizing that most vehicles have silhouettes that may be described as a “rounded rectangle”. The position and shape of such rounded rectangles is described in terms of 5 parameters (xc, yc, a, b, ?) where xc, yc are the center coordinates, a,b are the major and minor axis and ? is the orientation. The computation of a five dimensional Hough transform on an image including such a target will produce the five parameters that provide the “best fit” rounded rectangle to the target. These parameters are then passed to a missile tracker. This capability can be used to improve track gate handoff from the automatic target recognizer (ATR) to the missile tracker, changing aspect ratios of maneuvering targets, limited lock-on after launch (LOAL), aimpoint designation and fire control system to missile seeker handover.

Abstract: A non-traditional, non-uniformity compensation processor. In the illustrative embodiment, the feedforward adaptive non-uniformity compensation processor includes a multiplier which receives a video image and generates a compensated video image by multiplying the video signal by one of a noise reducing constant or one. The inventive processor further includes a shunting multiplication processor which supplies a selected one of the noise reducing constant or one to the multiplier in response to the presence of one of fixed pattern noise (FPN) and temporal noise (TN) in adjacent frames of the video image. In an illustrative embodiment, the inventive system cones the output of an FPA to generate a coned output signal. The system isolates the FPN in the coned output signal and feedforward processes the coned output signal on a pixel-by-pixel basis to generate a FPN-reduced output signal.

Abstract: A tracking system controls an aimed device (32) to keep it aimed at a target (34). A control system (10) determines when the tracker has lost its lock on the target (34) by comparing the target's instantaneous acceleration with its median acceleration. When the difference exceeds a predetermined threshold, the system searches backwards through a chronological buffer until it finds position data which antedate the receipt of inaccurate position data from the tracker. A second-order filter such as a Kalman filter is used to provide estimated target states. The difference between a time updated output from the second-order filter (16) and the present measured state from the tracking device is low pass filtered to provide a measurement of the instantaneous acceleration of the target. The buffer (20) stores position, velocity, and median acceleration values covering a span of time at least about twice as long as the time required to determine that a breaklock has occurred.

Abstract: A tracking system controls an aimed device (32) to keep it aimed at a target (34). A control system (10) determines when the tracker has lost its lock on the target (34) by comparing the target's instantaneous acceleration with its median acceleration. When the difference exceeds a predetermined threshold, the system searches backwards through a chronological buffer until it finds position data which antedate the receipt of inaccurate position data from the tracker. A second-order filter such as a Kalman filter is used to provide estimated target states. The difference between a time updated output from the second-order filter (16) and the present measured state from the tracking device is low pass filtered to provide a measurement of the instantaneous acceleration of the target. The buffer (20) stores position, velocity, and median acceleration values covering a span of time at least about twice as long as the time required to determine that a breaklock has occurred.

Abstract: A fuzzy controller (130) for use in target detection systems. The fuzzy controller (130) includes a first circuit (132) for determining a number of false alarms in a frame of data. A second circuit (136) determines a desired number of false alarms for the frame of data. A third circuit (134-152) computes a threshold multiplier factor based on the number of false alarms, the desired number of false alarms, and one or more fuzzy rules. In a specific embodiment, the first circuit (132) includes a target detection system for providing addresses of pixels whose values are within a predetermined range relative to a detection threshold. The second circuit (136) includes an input device (136) for accepting the desired number of false alarms as input to the fuzzy controller (130). The third circuit (138-152) includes a fuzzifier input calculation circuit (134) that computes a fuzzifier input value based on the number of false alarms detected in a frame and the desired number of false alarms.

Abstract: An accurate target detection system. The system includes a sensor (22) that receives electromagnetic signals and provides electrical signals in response thereto. A non-uniformity correction circuit (28, 38, 52) corrects non-uniformities in the sensor (22) based on the electrical signals and provides calibrated electrical signals in response thereto. A third circuit (30, 32, 34, 38, 42, 44, 52) determines if a target signal is present within the calibrated electrical signals and provides a target detection signal in response thereto. A fourth circuit (38, 40, 48) selectively activates or deactivates the non-uniformity correction circuit (28, 38, 52) based on the target detection signal.

Abstract: A system and method of finding a detection threshold that separates real scene objects seen through an optical system from sensor fixed artifacts. The inventive system continuously scans the sensor so that objects in the scene move in an inverse sense relative to the scanning motion while fixed frame artifacts remain fixed in focal plane coordinate system. The difference in temporal behavior is used to discriminate against fixed frame artifacts. The inventive system finds a detection threshold that maximizes the portion of a target that may be segmented while avoiding setting the threshold so low that it merges spatially adjacent artifacts in with the target.

Abstract: An accurate target detection system. The system includes a sensor (22) that receives electromagnetic signals and provides electrical signals in response thereto. A non-uniformity correction circuit (28, 38, 52) corrects non-uniformities in the sensor (22) based on the electrical signals and provides calibrated electrical signals in response thereto. A third circuit (30, 32, 34, 38, 42, 44, 52) determines if a target signal is present within the calibrated electrical signals and provides a target detection signal in response thereto. A fourth circuit (38, 40, 48) selectively activates or deactivates the non-uniformity correction circuit (28, 38, 52) based on the target detection signal. In a specific embodiment, the sensor (22) is an array of electromagnetic energy detectors (22), each detector providing an electrical detector output signal.

Abstract: A process for reducing background clutter of a scene, and isolating an object of interest. The process includes calculating a minimum output from a plurality of appropriately configured anti-median filters representing a minimum difference processor (MDP) filter. In a first embodiment, the MDP filter includes a (5.times.5) matrix of elements wherein each element has an intensity value representation of the scene. A first anti-median filter array is configured as a horizontal array of 5 pixel elements in which the center element of the matrix is the center element of the horizontal array. A median value of the 5 elements of the horizontal array is determined, and is subtracted from the value of the center element to determine an anti-median value of the horizontal array. The same process is performed for a vertical 5 element anti-median array, and two diagonal 5 element anti-median arrays, all including the center element of the matrix as the center common element of the arrays.

Abstract: A processor and method for discriminating against interference during target acquisition and reacquisition processing of densely cluttered images. The processor and method that adaptively estimates the feature probability density function of the interference from the image data. The estimated interference probability density function, along with target feature estimates are input to a Bayesian classifier that discriminates the interference from the target.

Abstract: Moving objects (36) are detected by processing images through a spatial filter (42) and positional shifter (46), which shifts the image by an amount reflecting a line of sight velocity hypothesis of the objects. The images are stacked together by a stacker 48, causing the intensity of an object moving at the line of sight velocity hypothesis to increase, and other objects to be blurred. The stacked image is spatially filtered (50) to remove the blurred objects and linear artifacts, and the moving objects (36) of interest are selected according to their increased intensities. In a practical system, the images are processed with a range of velocity hypotheses to identify both the object and the true line of sight velocity.

Abstract: A system for detecting an object in a scene. The invention (10) includes memory (12) for storing a first frame of image data represented by a plurality of input signals I.sub.o. A spatial filter (16) is included for spatially filtering a second frame of image data represented by a plurality of input signals I.sub.o to provide first spatial filter output signals of the form .eta..sub.s I.sub.o. These signals are subtracted from the first frame of image data to provide difference signals of the form .eta..sub.D I.sub.o, where .eta..sub.D represents a measure of object registration. A second spatial filter (18) is included for spatially filtering the difference signals to provide spatially filtered difference signals of the form .eta..sub.s .eta..sub.D I.sub.o. Finally, the signals are processed (30) to extract the measure of object registration .eta..sub.D.

Abstract: An improved method for finding the mask threshold for a maskable bilevel correlator operates on a series of frames of information. Beginning with an initial threshold, two rates are computed. The rate of change of the correlation peak magnitude P.sub.(T), a function of threshold T, with respect to the change in the value of the mask threshold T, or (dP.sub.(T) /dT), and the corresponding rate of change of correlation reference pixels with magnitude exceeding the mask threshold, U.sub.(T), a function of threshold T, with respect to the change in the value of the mask threshold T, or (dU.sub.(T) /dT), are measured. The correct, or best, value of the threshold occurs when .vertline.dP.sub.(T) /dT.vertline..gtoreq.0.5*.vertline.dU.sub.(T) /dT.vertline..

Abstract: A method and apparatus for detecting an object of interest against a cluttered background scene. In a first preferred embodiment the sensor tracking the scene is movable on a platform such that each frame of the video representation of the scene is aligned, i.e., appears at the same place in sensor coordinates. A current video frame of the scene is stored in a first frame storage device (14) and a previous video frame of the scene is stored in a second frame storage device (20). The frames are then subtracted by means of an invertor (24)and a frame adder (28) to remove most of the background clutter. The subtracted image is put through a first leakage reducing filter, preferably a minimum difference processor filter (32). The current video frame in the first frame storage device (14) is put through a second leakage-reducing filter, preferably minimum difference processor filter (36).

Abstract: Three image frames containing an object of interest and background clutter are taken at successive time intervals and stored in memory. The background of images A, B and C are registered preferably using an area correlator 12. A median filter 16 is used to select a median value from the registered image frames. Then, subtractor 18 serves to subtract the median pixel values from one of the image frames. This difference output is then thresholded to provide a binary signal whose pixel values exceeding the threshold levels are generally associated with the position of the moving object.

Abstract: Target tracking capability is provided by a synergistic tracker system (5) which includes both a correlation tracker (30) and an object tracker (40) for processing sensor data input (10) and for generating tracking error signals (80). The operation of the synergistic tracker system is controlled by a central processing unit (70). The system operates by first correlating a reference region image with a portion of a current digitized image provided by an analog to digital converter (20). Secondly, the object tracker provides a precisely defined trackpoint for an object within the current image. The correlation tracker stabilizes and limits the portion of the digitized image that the object tracker must operate upon. Stabilizing and limiting this portion of the digitized image reduces the object tracker's sensitivity to background clutter and sensitivity to a loss of lock induced by sensor motion. The object tracker provides a non-recursive update for the correlation's reference region image.

Abstract: A method and apparatus for correlating two signals representing a live image and a reference image are disclosed. The reference signal is processed to provide a polarity bit and mask bit for each pixel position in the reference image. The processed video signal is stored in a memory 20 and used in a convolver and summer section 18 in which the polarity bits between the reference and live images are gated by gate 17 and are used in generating a correlation output from summer 23. However, the polarity bits do not affect the correlation output if the logical output from gate 19 is of a given state which occurs if the associated mask bits indicate that the gradient value at the particular pixel position does not exceed certain minimum threshold values.

Abstract: A conventional tracker 14 is used to continually aim a sensor 12 at a flying object so that it is located in substantially the same spatial location in a plurality of image frames taken at different times. Thus, the object remains in about the same spatial position in each image frame while the background clutter generally changes. By processing two or more of the image frames it is possible to calculate a more precise aimpoint for the object than is generally provided by the track point.