Autonomous Target Recognition/Identification
DOD objectives
Autonomous Target Recognition/Identification
Autonomous Situational Awareness | Under any Environmental Conditions A new Digital Tunable Filter in hardware provides the ability to recognize targets, without false alarms, thousands of times more accurate and faster than the fastest and most powerful algorithms in existence. 25 Nano Seconds Per Pixel The power of the Human Neo Cortex with the ability to recognize target and objective thousands of times faster than the fastest and most powerful algorithms in existence. Never before have drones, observation aircraft and satellites have been this close to autonomy. With the ability to recognize, process and report with unparalleled detection power and accuracy, in a matter of a few system clock cycles, this system is light years ahead of computers and algorithms available to any designer.
|
Bandwidth Setting
There is no limitation to the spectral frequency spacing. Any frequency can be detected as long as the intensity values are pre-loaded in the filter memory.
The filtering is done on a pixel by pixel basis which offers far higher resolutions and noise isolation than current approaches. In short, a narrow band filter receives the digital values of eight IR channels and checks them to a preset values stored in memory. If all eight values are within a specified range of the preset values, a match is indicated. Because this approach does not require intensive calculations, filtering takes place at a rate of 20 nanoseconds per pixel or 20 milliseconds for a frame of 1,000,000 pixels.
Variations of this filter have a widespread use in detection's of moving cars in a parking lot, but they are not suited to track and count vehicles in a fast moving street or highway. Tracking non-rigid targets in low-resolution images has long been realized as a region based correspondence problem, in which each target is mapped from one frame to the next according to its position, dimension, color and other contextual information. When multiple targets exist and their dimensions are not negligible in comparison with their velocities, occlusion or grouping of these targets is a routine event. This brings about uncertainty for the tracking, because the contextual information is only available for the group and individual targets cannot be identified.
The filtering is done on a pixel by pixel basis which offers far higher resolutions and noise isolation than current approaches. In short, a narrow band filter receives the digital values of eight IR channels and checks them to a preset values stored in memory. If all eight values are within a specified range of the preset values, a match is indicated. Because this approach does not require intensive calculations, filtering takes place at a rate of 20 nanoseconds per pixel or 20 milliseconds for a frame of 1,000,000 pixels.
Variations of this filter have a widespread use in detection's of moving cars in a parking lot, but they are not suited to track and count vehicles in a fast moving street or highway. Tracking non-rigid targets in low-resolution images has long been realized as a region based correspondence problem, in which each target is mapped from one frame to the next according to its position, dimension, color and other contextual information. When multiple targets exist and their dimensions are not negligible in comparison with their velocities, occlusion or grouping of these targets is a routine event. This brings about uncertainty for the tracking, because the contextual information is only available for the group and individual targets cannot be identified.
New Technology in Calibration and filtering of Video Signals in Hardware
The new Digital Tunable Filter, Shown in US patent 8159568 by Ned M Ahdoot issued May 5, 2012) is for identifications of targets in an electromagnetic field. The operation of the Tunable Filter is very similar to our daily life of watching T.V. It behaves the same way as our eyes and brains detect targets, with identifying primary colors of a pixel (or group of pixels). There are no computation intensive calculations compared to the present technologies. The new filtering and identification capability can be used as an stand alone or it can lend itself to the some of the developed technologies of identification for better results.
|
Two Methods of Calibration using Hardware
The technology allows for two methods of calibration. One method is to calibrate based upon the calibration channel such as gray scale in visible light or utilizing one or more calibration channels for IR. The second method is to calibrate the data channels based upon the availability of the calibration data (from adverse environment conditions) to calibrate regular data channels.
Proposed Architecture
This proposal relates to a new and revolutionary technology (Figure 1) that is aimed to achieve DHS’s needs for identification and quantization of harmful agents. The simplicity of the concept will replace the present bulky and heavy equipment, with far more accurate results in shortest times. Figure 1 is hardware and firmware architecture of the proposed Technology. The technology is ideal and highly suitable to for small UAV’s, requiring very little power and weight.
Real Time Passive Detection and Identification of Organic and Inorganic Substances
The following analysis is for the concentration measurements of CWAs and TICs in a closed environment. The same methodology can be used, to remotely (open system) measure concentration of CO2 in the presence of other gases utilizing high resolution IR detection methodology described above. The preferred method for the measurement outlined in the paper is based upon a recent study and recommendation for measurements of chemical agents by, Michael B. Pushkarsky1, Michael E. Webber1, Tyson MacDonald1 and C. Kumar N. Patel and Department of Physics & Astronomy, University of California. Los Angeles, CA 90095.and US patent 8159568 by Ned M Ahdoot.
The UCLA study presents an analytical model for evaluating the suitability of optical absorption based spectroscopic techniques for detection of chemical warfare agents (CWAs) and toxic industrial chemicals (TICs) in a ambient air (closed system in which the sampled contaminated air is introduced to measurement system). The study is based upon sensor performances that are modeled by simulating absorption spectra of a sample containing both the target and multitude of interfering species as well as an appropriate stochastic noise for determining the target concentrations from the simulated spectra via a least square fit (LSF) algorithm. The distribution of the LSF target concentrations determines the sensor sensitivity, probability of false positives (PFP) and probability of false negatives (PFN). Their model was applied to CO2 laser based photoacosutic (L-PAS) CWA sensor and predicted single digit ppb sensitivity with very low PFP rates in the presence of significant amount of interferences. This approach will be useful for assessing sensor performance by developers and users alike; it also provides methodology for inter-comparison of different sensing technologies. Feasibility Discussions of Calibration This proposal utilizes the calculation methodology of the UCLA paper.
It need to be emphasized that the paper is for a closed system of measurements and use of lasers to achieve measurements of harmful gases in the presence of other gases. This paper seeks the feasibility of CO2 measurement, utilizing the stated and proven UCLA paper along with the very high resolutions and calibrations (discussed above) of this technology for remote sensing measurements. The essential reason and argument that can establish such identification and measurement remotely are: • Very high resolution capabilities of the new filtering technology for visible light and IR. Feasibility of providing much higher resolutions (compared to the proposed above), by increasing the number of channels and the number of bits in A/D converters. • Recommended approaches for correcting system imperfections and calibration due to adverse environment conditions. • Recommended approaches for calibration, even in the presence of adverse environmental conditions.
The UCLA study presents an analytical model for evaluating the suitability of optical absorption based spectroscopic techniques for detection of chemical warfare agents (CWAs) and toxic industrial chemicals (TICs) in a ambient air (closed system in which the sampled contaminated air is introduced to measurement system). The study is based upon sensor performances that are modeled by simulating absorption spectra of a sample containing both the target and multitude of interfering species as well as an appropriate stochastic noise for determining the target concentrations from the simulated spectra via a least square fit (LSF) algorithm. The distribution of the LSF target concentrations determines the sensor sensitivity, probability of false positives (PFP) and probability of false negatives (PFN). Their model was applied to CO2 laser based photoacosutic (L-PAS) CWA sensor and predicted single digit ppb sensitivity with very low PFP rates in the presence of significant amount of interferences. This approach will be useful for assessing sensor performance by developers and users alike; it also provides methodology for inter-comparison of different sensing technologies. Feasibility Discussions of Calibration This proposal utilizes the calculation methodology of the UCLA paper.
It need to be emphasized that the paper is for a closed system of measurements and use of lasers to achieve measurements of harmful gases in the presence of other gases. This paper seeks the feasibility of CO2 measurement, utilizing the stated and proven UCLA paper along with the very high resolutions and calibrations (discussed above) of this technology for remote sensing measurements. The essential reason and argument that can establish such identification and measurement remotely are: • Very high resolution capabilities of the new filtering technology for visible light and IR. Feasibility of providing much higher resolutions (compared to the proposed above), by increasing the number of channels and the number of bits in A/D converters. • Recommended approaches for correcting system imperfections and calibration due to adverse environment conditions. • Recommended approaches for calibration, even in the presence of adverse environmental conditions.
Measurements of harmful Chemical Intensities
IR absorption spectroscopy is a powerful tool for trace gas detection because a vast majority of polyatomic molecules including CWAs, TICs and explosives absorb light in the wavelength region from 3 to 14 um. Table 1 show the IR spectra of nerve and mustard gases and gives more comprehensive list of both CWAs and TICs absorbing between 3 and 11.5 um. The required sensitivity for CWA detection is determined by the toxicity levels of particular agents, most of which reasonably well documented. The concentrations and related health effects for Sarin (GB), a typical nerve agent CWA, are summarized in Table 1.
The characteristics that allows for the proposed system to measure chemical agents (from very long distances) are as follows:
The characteristics that allows for the proposed system to measure chemical agents (from very long distances) are as follows:
- The extremely high resolution in detection as explained above.
- The ability to automatically calibrate and correct the effects of distortions (pixel by pixel in real time), caused by the effects of adverse environment condition.
- The ability to make automatic adjustments of the system imperfections and the system temperature changes (Such as FPA) that affects measurement.
Method
The approach is based upon utilizing multispectral IR detection technique similar co2 lasers to achieve required sensitivity in the desired 9-11.5 nm spectral range. The following Figure is showing the absorption spectra of eight target nerve agents and surrogates at wavelengths accessible with close to 80 channel multispectral channel. Consequently 88 channels IR spectral from 9.6 to 11.4 micron were chosen as spectral range for the spectral libraries for DIMP all listed in above Table 1. Detections will produce the concentrations measurements by evaluating the absorption coefficients from the available libraries at 88 channel IR spectral wavelengths. |
Concentration measurements of Harmful Gases
In addition to sensitivity and fast response time, the very important parameter of a sensor is selectivity i.e. an ability to operate with acceptably low probability of false positive (PFP) readings in the presence of interferences typical in indoor and outdoor environments. In fact, the lack of selectivity may become a critical problem when sensing realistically contaminated air because of high economic costs of frequent false alarms. Therefore a model allowing prediction of the sensor PFP as well as the probability of false negative (PFN) would be very valuable tool for sensor developers and users alike. This paper presents an analytical model allowing to quantitatively evaluating performance, i.e. sensitivity and selectivity of absorption spectroscopy based techniques for the detection of CWAs and TICs in realistically contaminated ambient air.
The concentration measurements equations are given by:
The above equation provides needed mathematical tools for concentration measurements of multi sensor which relies upon acquisition and analysis of a complex spectrum. For practical application of the presented model one needs to specify the following parameters:
- A set of concentrations, X i , of all expected targets and interferences
- The spectral range of a spectrometer and the absorption spectral library of all the expected ambient air constituents that absorb appreciably within an identified spectral range
CO2 Measurements
Calibration and filtering of video signals will allow effective detections and concentration measurements of CO2 in minimum amount of time. Remote sensing of harmful chemicals is best accomplished by using visible light and IR spectrum, Detection, tracking, classifying, and responding to all chemical and biological weapons requires a new remote detection technology that provides rapid calibration of detected signals against the numerous adverse environmental conditions as well as very high degree of identification and measurements, even in the presence of other gas agents. The two characteristic increases the Probability of Detection (PD) and prevention of false alarms and ability of using fusion methods that receives data from many sensors. The new technology should be able to:
- Provide situational awareness instantly to an operator in relation to a certain geographic area.
- Operate at low power consumption levels
- Space and Ground based identification and tracking
- A fast system operates at real time that is deployed in any aerial systems with the ability to provide ground operators with situational awareness