Fig 4 - uploaded by Paul F. Mcmanamon
Content may be subject to copyright.
Reflectance spectra in the 1 to 5 m wavelength range for the four material classes studied in Section 6.
Source publication
A statistical spectral band selection procedure and classifiers for an active multispectral laser radar (LADAR) sensor are described. The sensor will operate in the 1 to 5 m wavelength region. The algorithms proposed are tested using library reflectance spectra for some representative background materials. The material classes considered include bo...
Contexts in source publication
Context 1
... demonstrate the band selection procedure described in Section 4 and the classifiers described in Section 5, we present some simulation results here. For this analysis, we have selected four classes of materials with the spectral reflectances shown in Fig. 4. These data include 210 spec- tral samples over the wavelength range of 1 to 5 m. At- mospheric corrections are then applied to these spectra. The atmospheric attenuation results from a variety of factors including molecular absorption, scattering, fog, precipita- tion, dust, etc. Several useful atmospheric models such as moderate ...
Context 2
... transmis- sion HITRAN have been developed at Air Force Phillips Laboratories, Hanscom Air Force Base. Here, MODTRAN is used and provides a wavenumber spectral resolution of 2 cm 1 for molecular and aerosol constituents. The MODTRAN atmospheric transmission data were resampled to match the wavelengths of the library reflectance spectra shown in Fig. 4. The computed one-way atmospheric trans- mission spectra in the 1 to 5 m wavelength range are shown in Fig. 5 for two different meteorological visibility conditions. The class spectra with two-way atmospheric correction range 10 km, altitude of 15,000 ft and visibility of 23 km are shown in Fig. ...
Context 3
... the four classes of materials, shown in Fig. 4, arranged according to the spatial template shown in Fig. 10a. Each class is coded with a different gray level. Using the atmospherically corrected reflectance spectra, shown in Fig. 6, three spectral band images were simulated. These are shown in Figs. 10b to 10d at the optimal wave- lengths calculated in the previous section. Again, ...
Similar publications
Metasurface yields fantastic progress in diverse functional electromagnetic devices. However, in previously reported works, high-performance camouflaging device with curved metasurface for complex conformal target still keeps elusive. Herein, a novel and efficient Bezier-shaped metasurface is proposed based on the revised phase distribution equatio...
Citations
... Upgrading the healthcare database improves the heart of medical analysis [14]. Data pre-preparing steps are data integration, data cleaning, data reduction and data transformation that is feature subset selection [15][16][17][18][19][20]. A few ascribes of datasets might be repetitive as their data might be accommodated in other attributes. ...
Feature selection technique is an important step in the prediction and classification process, primarily in data mining related aspects or related to medical field. Feature selection is immersive with the errand of choosing a subset of applicable features that could be utilized in developing a prototype. Medical datasets are huge in size; hence some effective optimization techniques are required to produce accurate results. Optimization algorithms are a critical function in medical data mining particularly in identifying diseases since it offers excellent effectiveness in minimum computational expense and time. The classification algorithms also produce superior outcomes when an objective function is built using the feature selection algorithm. The solitary motive of the research paper analysis is to comprehend the reach and utility of optimization algorithms such as the Genetic Algorithm (GA), the Particle Swarm Optimization (PSO) and the Ant Colony Optimization (ACO) in the field of Health care. The aim is to bring efficiency and maximum optimization in the health care sector using the vast information that is already available related to these fields. With the help of data sets that are available in the health care analysis, our focus is to extract the most important features using optimization techniques and work on different algorithms so as to get the most optimized result. Precision largely depends on usefulness of features that are taken into consideration along with finding useful patterns in those features to characterize the main problem. The Performance of the optimized algorithm finds the overall optimum with less function evaluation. The principle target of this examination is to optimize feature selection technique to bring an optimized and efficient model to cater to various health issues. In this research paper, to do bibliometric analysis Scopus and Web of Science databases are used. This bibliometric analysis considers important keywords, datasets, significance of the considered research papers. It also gives details about types, sources of publications, yearly publication trends, significant countries from Scopus and Web of Science. Also, it captures details about co-appearing keywords, authors, source titles through networked diagrams. In a way, this research paper can be useful to researchers who want to contribute in the area of feature selection and optimization in healthcare. From this research paper it is observed that there is a lot scope for research for the considered research area. This kind of research will also be helpful for analyzing pandemic scenarios like COVID-19.
... Most of these sensors are used for man-made objects and environments, as it is a bit complicated to use such sensors in natural objects, especially trees and other vegetation sometimes occlude objects of interest. For some applications, methods have been developed to filter such unwanted objects from LIDAR scans to gain a better view of objects of interest [12]. ...
Proposed is a minimal setup for proving feasibility of using a very cost-effective LIDAR sensor to be used in advanced applications where high-end LIDAR sensors are normally used. Here the application demonstrated is with low computing resources and low-cost devices for demonstrating 2D mapping of a Horticulture field. The setup proposed has got an overall Error of 1.08% while the sensor itself had 1% error. Hence it can be approximated that the setup error is 0.08%.
... Wavelength diversity is available if you are willing to provide laser illuminators at multiple wavelengths, and the associated detectors. [79][80][81][82][83][84] Fig. 16 Difference between real beam resolution and flash aperture synthesis resolution. Optical Engineering 060901-11 June 2012/Vol. ...
Ladar is becoming more prominent due to the maturation of its component
technologies, especially lasers. There are many forms of ladar. There is
simple two-dimensional (2-D) ladar, similar to a passive electro-optic
sensor, but with controlled illumination and the ability to see at night
even at short wavelengths. There is three-dimensional (3-D) ladar, with
angle/angle/range information. 3-D images are very powerful because
shape is an invariant. 3-D images can be easily rotated to various
perspectives. You can add gray scale or color, just like passive, or 2-D
ladar, imaging. You can add precise velocity measurement, including
vibrations. Ladar generates orders of magnitude higher frequency change
then microwave radar for velocity measurement, because frequency change
is proportional to one over the wavelength. Orders of magnitude higher
frequency change means you can measure a given velocity orders of
magnitude quicker, in many cases making an accurate measurement
possible. Polarization can be used. With an active sensor you control
both the illumination and the reception, so you can pattern the
illumination. Also, because ladar can use narrow band illumination it is
easier to easier to coherently combine sub-aperture images to obtain the
higher resolution of an array.
... Unlike other sensors that combine 3D and spectral imaging, this sensor computes a range for each spectral band independent of backscatter returns in all other bands. The technique and capabilities presented here contrast favorably with similar reports of experimental demonstrations, having higher range resolution [12,13], operating at longer maximum range [12,14], offering the possibility of very high pixel rates [15,16,17], implementing many (25) spectral bands over a broad optical bandwidth [13,14,18,19,20], using wavelengths considered to be eye safe [14,21] and realizing the advantage of physical-level combination in lieu of fusion of data from independent sensors to associate 3D points with spectra [20,21]. ...
We present the concept and experimental results for Spectral LADAR, an augmented LADAR imager combining three-dimensional (3D) time-of-flight ranging with active multispectral sensing in the shortwave infrared (1080–1620 nm). The demonstrated technique is based on a nanosecond regime pulsed supercontinuum transmitter and spectrally multiplexed receiver that computes a high-resolution range value for each of 25 spectral bands. A low frame-rate prototype unit is described. Results demonstrating 3D imaging and material type classification of objects, especially those obscured by camouflage, are shown at effective stand-off ranges exceeding 40 m. These capabilities and the highly eye safe wavelengths at which the system operates make it suitable for applications in military imaging and robotic perception.
... Natural objects, especially trees and other vegetation, sometimes occlude objects or scenes of interest; thus, they are ignored or avoided. In some cases methods have been developed to filter vegetation and other natural materials from lidar scenes to gain a better view of objects of interest (Hardie et al. 1998). ...
A commercially available pan-and-tilt mount laser scanner was used to acquire data for subsequent three-dimensional modeling and measurement of standing forest trees. Methods were developed for identifying trees in range images and co-registering range images acquired from different vantage points. Upper-stem diameters and branch heights derived from the range images were compared to measurements made after the felling of a small number of loblolly pine (Pinus taeda L.) trees. Tree identification assumed bole cross-sections were circular, estimating their geometric centers at successive heights up the stem. Tree center estimates at multiple heights were then used to co-register images made from different vantage points. Co-registration (x, y) errors did not exceed 2.1 cm in any of the 18 pairwise registrations carried out. Results showed excellent agreement (average error < 1 cm) between the lidar-derived diameter estimates and caliper measurements for bole sections below the base of live crown. Less a
... In order to maximize the performance of the band optimization algorithm and the material classification system, the probability of error for the selected classifier and observation model must be minimized. In a two-class system with Gaussian power density functions, computing the probability of error is unwieldy and is usually replaced by applying a bound metric such as the Bhattacharyya bound [13]. In multi-class scenarios, no such bound exists and forces the use of a more practical approach, such as the aforementioned class separability matrix. ...
... where tr() represents the trace operation [13]. Large values of Jindicate large betweenclass scatter separation relative to within-class scatter. ...
... The average transmission coefficients for this scenario are 40.7% and 92.7%. These values compared to the previous case fall in line13 The absorption band centered at 2.7u.m is an aggregate of C0 2 and H 2 0 molecular absorption. ...
“Active” sensors (systems that employ lasers as an illumination source) have several advantages. In particular, multiple phenomenologies can be measured. This paper describes the types of measurements possible and the required source and receiver components.
Active imaging and mapping using lasers as illumination sources have been of increasing interest during the last decades. Applications range from defense and security, remote sensing, medicine, robotics, and others. So far, these laser systems have mostly been based on a fix wavelength laser. Recent advances in lasers enable emission of tunable, multiline, or broadband emission, which together with the development of array detectors will extend the capabilities of active imaging and mapping. This paper will review some of the recent work on active imaging mainly for defense and security and remote sensing applications. A short survey of basic lidar relations and present fix wavelength laser systems is followed by a review of the benefits of adding the spectral dimension to active and/or passive electro-optical systems.
We present a new technique for adaptive band selection from hyperspectral image cubes for detecting small targets using an anomaly detector. The proposed technique ensures the selection of lowest number of spectral bands using Mahalanobis distance, maximum affordable extra noise variance, and Constant False Alarm Rate (CFAR) anomaly detector threshold. Since the target is small, band selection based only on Mahalanobis distance is not adequate and the bands which can withstand extra noise should be chosen. In addition, thresholding must be considered to avoid misclassification of target. The proposed technique yields comparatively low false alarm rate even with a few bands. For real time applications, acousto-optic tunable filter based spectral imagers may be used for acquiring hyperspectral images and for selecting the bands.
A comparison of the spectral bands recommended through employment of different data separation measures and the reliability and robustness of these measures was performed on artificially generated target and background IR radiance data sets. The Mahalanobis distance, Signal to Clutter Ratio, Bhattacharya distance and Informational Difference criteria were employed in order to obtain the best single and paired spectral bands for data separation between two data classes of 'targets' and 'backgrounds' in day and night conditions. The results show that for conditions in which there is a distinct temperature difference between the two data classes, all the criteria perform similarly, with only small differences in the recommended spectral bands and general performance. However, in daylight conditions with multiple types of backgrounds and targets, criteria based on the assumption of concentrated data classes (SCR, Mahalanobis) tend to provide contradictory results, while those based on general statistical principles (Bhattacharya, Informational Difference) produce unequivocal results that are relatively unaffected by data set complexity.
