Thermal imagers and used therein infrared array sensors are subject to calibration procedure and evaluation of their
voltage sensitivity on incident radiation during manufacturing process. The calibration procedure is especially important
in so-called radiometric cameras, where accurate radiometric quantities, given in physical units, are of concern. Even
though non-radiometric cameras are not expected to stand up to such elevated standards, it is still important, that the
image faithfully represents temperature variations across the scene. The detectors used in thermal camera are illuminated
by infrared radiation transmitted through a specialized optical system. Each optical system used influences irradiation
distribution across an sensor array. In the article a model describing irradiation distribution across an array sensor
working with an optical system used in the calibration set-up has been proposed. In the said method optical and
geometrical considerations of the array set-up have been taken into account. By means of Monte-Carlo simulation, large
number of rays has been traced to the sensor plane, what allowed to determine the irradiation distribution across the
image plane for different aperture limiting configurations. Simulated results have been confronted with proposed
analytical expression. Presented radiometric model allows fast and accurate non-uniformity correction to be carried out.
Contemporary infrared detector arrays suffers from technological imprecision which causes that the response to uniform radiation results in nonuniform image with superimposed fixed pattern noise (FPN). In order to compensate this noise there is a need to evaluate detectors characteristics like responsivity and offset of every detector in array. In article the method of determining the responsivity of detectors in a microbolometer array is described. In the method geometrical and optical parameters of the detector array and the measurement system are taken into account. A special test bench was constructed and is consisting of: two precise surface black bodies, aperture limiter, an electronic interface for data acquisition and software for measurement and correction of results with optical parameters of the measuring stand taken into account. Constructed aperture limiter enables evaluation of optical paths in measurement stand with equivalent relative aperture F# from 0.5 to 16.1 In order to evaluate the impact of optical path to radiation distribution in the measurement system, special radiation model was elaborated and evaluated in Zemax software. Incident radiation intensity distribution on the detector surface was calculated using Monte-Carlo method for various parameters of the optical path in the measurement system. Calculated radiation maps were used to compensate radiation intensity nonuniformity of optical measurement system giving more precise responsivity evaluation of detector array parameters. The obtained values of voltage responsivity of the detectors in the array, can be used in algorithms like nonuniformity correction and radiometric calibration of the infrared camera. In article results of responsivity evaluation is presented for microbolometer infrared arrays from ULIS company (France).
The interpretation of IR images depends on radiative properties of observed objects and surrounding scenery. Skills and experience of an observer itself are also of great importance. The solution to improve the effectiveness of observation is utilization of algorithm of image enhancing capable to improve the image quality and the same effectiveness of object detection. The paper presents results of testing the hardware implementation of IR image enhancing algorithm based on histogram processing. Main issue in hardware implementation of complex procedures for image enhancing algorithms is high computational cost. As a result implementation of complex algorithms using general purpose processors and software usually does not bring satisfactory results. Because of high efficiency requirements and the need of parallel operation, the ALTERA’s EP2C35F672 FPGA device was used. It provides sufficient processing speed combined with relatively low power consumption. A digital image processing and control module was designed and constructed around two main integrated circuits: a FPGA device and a microcontroller. Programmable FPGA device performs image data processing operations which requires considerable computing power. It also generates the control signals for array readout, performs NUC correction and bad pixel mapping, generates the control signals for display module and finally executes complex image processing algorithms. Implemented adaptive algorithm is based on plateau histogram equalization. Tests were performed on real IR images of different types of objects registered in different spectral bands. The simulations and laboratory experiments proved the correct operation of the designed system in executing the sophisticated image enhancement.
The paper presents the thermal sight for small arms weapons, which can be classified as 3rd gen thermal camera. The
sight operates in LWIR (long wave infrared) spectra band and utilizes uncooled microbolometer focal plane array (FPA)
with stabilized temperature (by means of Peltier module). The assumed technical and tactical characteristics of the
presented sight were confirmed during laboratory test (including climate and vibration tests). The sight was also tested
during field trials conducted at Military Institute of Armament Technology, where it was mounted on seven different
weapon types with calibers from 5.56 to 12.7 mm.
A modification of Sum-of-Squared-Differences algorithm is proposed to improve tracking efficiency of small objects in
infra-red image sequences. The reason to use SSD algorithm is its better performance in tracking small objects, than in
model based tracking algorithms. However traditional Sum-of-Squared-Differences (SSD) algorithm is sensitive to
partial or full occlusions, background clutter and changes in object appearance. To increase immunity to this kind of
noises the modification in model update procedure was developed. The experimental results illustrate that the proposed
modification to SSD algorithm can improve overall algorithm performance in infrared operation. The paper describes the
Sum-of-Squared Differences algorithm and its principal features in tracking objects on thermal image sequences. Next
modification to SSD algorithm is described. Finally the experimental results are presented with comparison between
traditional and modified SSD algorithm.
The paper discusses technical possibilities to build an effective electro-optical sensor unit for sniper detection using
infrared cameras. This unit, comprising of thermal and daylight cameras, can operate as a standalone device but its
primary application is a multi-sensor sniper and shot detection system. At first, the analysis was presented of three
distinguished phases of sniper activity: before, during and after the shot. On the basis of experimental data the
parameters defining the relevant sniper signatures were determined which are essential in assessing the capability of
infrared camera to detect sniper activity. A sniper body and muzzle flash were analyzed as targets and the descriptions of
phenomena which make it possible to detect sniper activities in infrared spectra as well as analysis of physical limitations
were performed. The analyzed infrared systems were simulated using NVTherm software. The calculations for several
cameras, equipped with different lenses and detector types were performed. The simulation of detection ranges was
performed for the selected scenarios of sniper detection tasks. After the analysis of simulation results, the technical
specifications of infrared sniper detection system were discussed, required to provide assumed detection range. Finally
the infrared camera setup was proposed which can detected sniper from 1000 meters range.
Infrared cameras are used in various military applications for early detection and observation. In applications where very
fast image acquisition is needed the so called cooled detectors are used. Cooled detectors are a kind of detectors that
demands cryogenic cooling, but in return provide exceptional performance and temperature sensitivity with low
integration times. These features predestinate cooled detectors for special purposes like airborne systems, where fast and
precise infrared radiation measurement is needed. Modern infrared cooled detector arrays like HgCdTe Epsilon detector
from Sofradir with spectral range of 3.5μm-5μm can provide high frame rate reaching 140Hz with full frame readout.
Increasing frame rates of cooled infrared detectors demands fast and efficient image processing modules for necessary
operations like nonuniformity correction, bad pixel replacement and visualization. For that kind of detector array a fast
image processing module was developed.
The module is made of two separate FPGA modules and configuration processor. One FPGA was responsible for
infrared data processing, and was performing nonuniformity correction, bad pixel replacement, linear and nonlinear
filtering in spatial domain and dynamic range compression. Second FPGA was responsible for interfacing infrared data
stream to standard video interfaces. It was responsible for frame rate conversion, image scaling and interpolation, and
controlling ASICs for video interface realization. Both FPGAs use several external resources like SRAM and DRAM
memories. The input interface was developed to connect with Epsilink board which is a standard proximity board
provided by Sofradir for this kind of detector. The image processing chain is capable of performing real-time processing
on data stream of volume up to about 40 Megapixels per second.
Images produced by IR cameras are a specific source of information. The perception and interpretation of such image
greatly depends on thermal properties of observed object and surrounding scenery. In practice, the optimal settings of the
camera as well as automatic temperature range control do not guarantee the displayed images is optimal from observer's
point of view. The solution to this could be the methods and algorithms of digital image processing implemented in the
camera. Such solution should provide intelligent, dynamic contrast control applied not only across entire image but also
selectively to specific areas in order to maintain optimal visualization of observed scenery. The paper discusses problems
dealing with improvement of the visibility of low-contrast objects and presents method of image enhancement. The
algorithm is based on adaptive histogram equalization. The image enhancement algorithm was tested on real IR images.
The algorithm significantly improves the image quality and the effectiveness of object detection for the majority of
thermal images. Due to its adaptive nature it should be effective for any given thermal image. The application of such
algorithm is promising alternative to more expensive opto-electronic components like improved optics and detectors.
In article a digital system for high resolution infrared camera control and image processing is described. The camera is
built with use of bolometric focal plane array of size 640 by 480 detectors. Single detector in array has size of 25 μm and
can detect incident radiation from the spectral range of 8÷12 μm thanks to the special filter installed in specially designed
entrance window. The most important tasks of infrared image processing system are array readout and correction of
detectors offset and responsivity variations. The next tasks of the system are conversion of analog voltage signals from
microbolometers in array to digital form and then composition of a thermal image. Microbolometer array needs to be
controlled via several signals. The signal generator for readout circuit is capable of changing various timing parameters
like frame rate or integration time of the detector array. The changes in these parameters can be done via special set of
memory mapped registers. The infrared data received from detector array is transferred via data bus to modules
performing image processing, for example techniques for image enhancement. Image processing algorithms necessary
for infrared image generation are nonuniformity correction, bad pixel replacement and radiometric calibration.
Optionally an additional image processing techniques can be performed like edge enhancement, dynamic range
compression or object identification. The elaborated architecture of the system allowed easy change of parameters of the
system and to adopt many new algorithms without significant hardware changes. Scientific work funded from science
fund for years 2009-2011 as a development project.
Rapid development of infrared detector arrays caused a need to develop robust signal processing chain able to perform
operations on infrared image in real-time. Every infrared detector array suffers from so-called nonuniformity, which has
to be digitally compensated by the internal circuits of the camera. Digital circuit also has to detect and replace signal
from damaged detectors. At the end the image has to be prepared for display on external display unit. For the best
comfort of viewing the delay between registering the infrared image and displaying it should be as short as possible. That
is why the image processing has to be done with minimum latency. This demand enforces to use special processing
techniques like pipelining and parallel processing.
Designed infrared processing module is able to perform standard operations on infrared image with very low latency.
Additionally modular design and defined data bus allows easy expansion of the signal processing chain. Presented image
processing module was used in two camera designs based on uncooled microbolometric detector array form ULIS and
cooled photon detector from Sofradir. The image processing module was implemented in FPGA structure and worked
with external ARM processor for control and coprocessing. The paper describes the design of the processing unit, results
of image processing, and parameters of module like power consumption and hardware utilization.
In areas like military systems, surveillance systems, or industrial process control, more and more often there is a need to
operate in limited visibility conditions or even in complete darkness. In such conditions vision systems can benefit by
using thermal vision cameras. In thermal imaging an infrared radiation detector arrays are used. Contemporary infrared
detector arrays suffers from technological imprecision which causes that the response to uniform radiation results in
nonuniform image with superimposed fixed pattern noise (FPN). In order to compensate this noise there is a need to
evaluate detectors characteristics like responsivity and offset of every detector in array. Some of the detectors in cooled
detector arrays can be also defective. Signal from defective pixels has to be in such system replaced. In order to replace
defective pixels, there is a need to detect them. Identification of so-called blinking pixels needs long time measurement,
which in designed calibration stand is also possible. The paper presents the design of infrared detector array
measurement stand allowing measurement of mentioned parameters. Measurement stand was also used to evaluate
temporal noise of infrared detection modules. In article there is a description of optical system design and parameters of
used reference blackbodies. To capture images from camera modules a specially designed digital image interface was
used. Measurement control and calculations were made in specially written IRDiag software. Stand was used to measure
parameters for cameras based on cooled focal plane arrays from Sofradir. Results of two-point nonuniformity correction
are also presented.
A microbolometer is an uncooled thermal sensor of infra-red radiation. In thermal imaging, microbolometers organized
in arrays called focal plane arrays (FPA) are used. Because of technological process microbolometric FPAs features
unwanted detector gain and offset nonuniformity. Because of that, the detector matrix, being exposed to uniform infrared
radiation produces nonuniform image with superimposed fixed pattern noise (FPN). To eliminate FPN, nonuniformity
correction (NUC) algorithms are used. The offset of detector in array depends from mean temperature of FPA. Every
single detector in matrix has its temperature drift, so the characteristic of every detector changes over temperature. To
overpass this problem, a temperature stabilization of FPA is commonly used, however temperature stabilization is a
relatively power demanding process. In this article a method of offset calculation and correction for every detector in
array in function of mean array temperature is described. The method of offset temperature characteristic estimation is
shown. The elaborated method let to use unstabilized microbolometric focal plane array in thermographic camera. Method of offset correction was evaluated for amorphous silicon based UL 03 04 1 detector array form ULIS.
Introduction of a ground multispectral detection has changed organization and construction of perimeter security
systems. The perimeter systems with linear zone sensors and cables have been replaced with a point arrangement of
sensors with multispectral detection. Such multispectral sensors generally consist of an active ground radar, which scans
the protected area with microwaves or millimeter waves, a thermal camera, which detects temperature contrast and a
visible range camera. Connection of these three different technologies into one system requires methodology for
selection of technical conditions of installation and parameters of sensors. This procedure enables us to construct a
system with correlated range, resolution, field of view and object identification. The second technical problem connected
with the multispectral system is its software, which helps couple the radar with the cameras. This software can be used
for automatic focusing of cameras, automatic guiding cameras to an object detected by the radar, tracking of the object
and localization of the object on the digital map as well as identification and alarming.
In this paper two essential issues connected with multispectral system are described. We focus on methodology of
selection of sensors parameters. We present usage of a spider-chart, which was adopted to the proposed methodology.
Next, we describe methodology of automation of the system regarding an object detection, tracking, identification,
localization and alarming.
The methods of detection and identification of objects based on acoustic signal analysis are used in many applications, e.g., alarm systems, military battlefield reconnaissance systems, intelligent ammunition, and others. The construction of technical objects such as vehicle or helicopter gives some possibilities to identify them on the basis of acoustic signals generated by those objects. In this paper a method of automatic detection, classification and identification of military vehicles and helicopters using a digital analysis of acoustic signals is presented. The method offers a relatively high probability of object detection in attendance of other disturbing acoustic signals. Moreover, it provides low probability of false classification and identification of object. The application of this method to acoustic sensor for the anti-helicopter mine is also presented.
Proc. SPIE. 7113, Electro-Optical and Infrared Systems: Technology and Applications V
KEYWORDS: Infrared detectors, Signal to noise ratio, Infrared sensors, Sun, Detection and tracking algorithms, Sensors, Interference (communication), Signal processing, Signal analysis, Signal detection
PIR detectors used in security systems for people detection operate in far IR range (8÷14) mm. These detectors most frequently employ pyroelectric sensors. Application of a single pyroelectric sensor does not ensure distinguishing the phenomena of alarm character from, so-called, false alarms caused by, e.g., air turbulences or changes in a background temperature resulting from sun radiation. Thus, in PIR detectors, the sensors with two active elements are used (two sensors) and an alarm signal is determined on the basis of analysis of a difference (or a sum) and their output signals. Essential drawback of currently available PIR detectors is low efficiency of detection of slowly moving or crawling people. Efficiency of detection of slowly moving objects is low because radiation from such objects is close to background thermal noises.
The presented signal analysis is based on determination of average moving value in three "time windows" of a defined wavelength. Moreover, a principle of "time windows" creation is given and an algorithm for determination of detection thresholds is described. In PIR detector, an adaptation detection threshold was taken following thermal changes of a background. Influence of sun radiation is taken into account in the algorithm of determination of adaptation detection threshold.
The paper presents design and principle of operation of a passive IR detector of large detection range. Significant virtue of the described PIR detector is highly efficient detection of very slowly moving or crawling people. High signal-to-noise ratio was obtained by using larger number of pyroelectric sensors or by increasing number of detection zones (channels). Larger number of pyroelectric sensors forces development of a complex optical system. The presented optical system of PIR detector consists of one lens (germanium or amtir) and mirror concentrators. The optical system ensures continuity of detection zones (no "blind" area between particular detection zones).
Original electronic system for PIR detector was described in which direct current amplifiers of a signal from pyroelectric sensors were applied. Electronic system automatically reduces a voltage drift from pyroelectric sensors, thus significantly decreases low limit frequency of a conduction band of amplification channel. Together with a fulfillment of this condition, low-frequency noises enhancement is observed and next detector sensitivity diminishes. To ensure large detection ranges, a new method of signals analysis was applied.
PIR detector has been equipped with a channel of RS 485 standard data transmission. For registration of measurement results, special software was developed for detector diagnostics allowing registration of signals from particular detection zones. The investigation results for various ranges of PIR detector were presented. The signals from PIR detector were shown which were caused by crawling people being at the distance of 140 meters and walking, running people being at the distance more than 200 meters.