The Development of a 3D LADAR Simulator Based on a Fast Target Impulse Response Generation Approach
 467 Downloads
Abstract
A new laser detection and ranging (LADAR) simulator has been developed, using MATLAB and its graphical user interface, to simulate direct detection time of flight LADAR systems, and to produce 3D simulated scanning images under a wide variety of conditions. This simulator models each stage from the laser source to data generation and can be considered as an efficient simulation tool to use when developing LADAR systems and their data processing algorithms. The novel approach proposed for this simulator is to generate the actual target impulse response. This approach is fast and able to deal with high scanning requirements without losing the fidelity that accompanies increments in speed. This leads to a more efficient LADAR simulator and opens up the possibility for simulating LADAR beam propagation more accurately by using a large number of laser footprint samples. The approach is to select only the parts of the target that lie in the laser beam angular field by mathematically deriving the required equations and calculating the target angular ranges. The performance of the new simulator has been evaluated under different scanning conditions, the results showing significant increments in processing speeds in comparison to conventional approaches, which are also used in this study as a point of comparison for the results. The results also show the simulator’s ability to simulate phenomena related to the scanning process, for example, type of noise, scanning resolution and laser beam width.
Keywords
3D laser radar 3D LADAR simulator 3D LIDAR simulator 3D laser imaging1 Introduction
Laser detection and ranging, or laser radar (LADAR) systems, are considered an attractive alternative to radio detection and ranging (RADAR) systems because they use laser wavelengths which are shorter than RADAR wavelengths, to produces very highresolution 3D images. In addition, light velocity allows LADAR systems to take numerous measurements per second. LADAR images are created by scanning a scene with laser beams, the return time for these beams used to calculate range LADAR data. The format for this data is range, azimuth and elevation angle, this representing the spherical coordinates system whose origin is the sensor. LADAR converts this type of data into a 3D Cartesian format in order to produce a threedimensional range image, this in turn representing the spatial location of the intersection of the laser beam with the scanned scene.
LADAR systems play diverse roles in both civilian and military applications. In ground navigation, they are used for obstacle and roadboundary detection, and autonomous vehicle navigation [12, 14, 19, 29, 31, 32, 44, 47]. In aerial navigation, they provide autonomous navigational capacities [16], obstacle warning systems [11], and considered as a reliable alternative to GPS [40]. Regarding maritime navigation, LADAR are used for both precise manoeuvring operations and obstacle avoidance [24]. Looking to their use by the military, LADAR assists target detection and classification [8, 10, 39], antiship missile tracking [33], target identification at long range [5, 6, 7, 25, 41, 42], and the identification of military ground vehicles that may be hidden under camouflage or foliage such as tree canopies [30].
In consequence, simulations for LADAR systems have become a valuable tool for developing said systems and their data processing algorithms [13] because the simulator is able to produced LADAR images under different controlled effects, which enable the algorithms’ developers to evaluate their algorithms under these effects individually. In order to simulate these systems, a target impulse response must be generated for each laser pulse transmitted to the components of the target. This is an extensive computational process as the intersection points of each laser beam with the target’s surface, need to be identified in addition to their traveling distances.
Since most applications related to developing LADAR systems require LADAR simulators to be able to accurately simulate the propagation of the laser beam very quickly; this implies the need for rapid processing of a large number of laser footprint samples. In addition, in order to develop LADAR processing algorithms using simulated data, the simulators must be able to scan a large number of targets at high speed, under different scanning parameters.
Several methods have been developed to increase the speed of the required computational process; some approximate the simulation by defining the reflection of the laser pulse as 3D model voxels that have a direct lineofsight to the sensor [20, 21]. Others calculate the distance between the target and the sensor by using the division of the model’s surface and the distance between the viewpoint and model’s 3D points for simplification [45, 46]. A single wide laser beam projection, with focalplane array and parallel computing, is also used to reduce computational time [15, 22, 26, 43].
In this paper, a new approach to generate the actual target impulse response, based on finding the actual intersection points between the laser beam and the 3D model, is presented. This approach is based on deriving the algorithms required to calculate target angular ranges, these algorithms used to speed up the process. In order to evaluate the performance of the simulator using this approach, over forty 3D models were scanned, under different scanning parameters, the simulation times recorded.
In the following sections, the theoretical background, which includes the equations and parameters required to simulate the laser beam propagation and thus the core of the simulator, are presented. This is followed by the new approach of generating the target impulse response. The main concepts of the simulation implementation for the LADAR simulator and its control windows are described using a collection of selected simulated images. The testing procedure and results of the evaluation are given followed by the conclusion.
2 Simulation of Laser Beam Propagation
2.1 Laser Beam Energy Distribution
2.1.1 Temporal Distribution
2.1.2 Spatial Distribution
2.2 Atmospheric Effects
2.3 Target Interaction
The interaction between the transmitted laser beam and the target surface produces a reflected signal. The characteristics of this signal depend on the surface reflectance \(\rho _{tr}\) (\(225\%\)) [34], the angle of dispersion \(\varOmega _{tr}\) (Lambertian targets are assumed i.e. \(\varOmega _{tr}=\pi\) [9, 23, 34]), surface area \(A_{tr}\) (extended targets are assumed), surface shape and beam incidence angle.
2.4 LADAR Receiver
The process of determining the range to the target from the reflected signal is accomplished by the LADAR receiver. This process depends on detection techniques [3] (direct or coherent), optical transmission \(T_{o}\) (the fraction of energy that arrives at the detector from the total energy captured by the receiver aperture), quantum efficiency \(\eta\) (the fraction of the signal that is converted into photoelectrons) of the detector, pulse detection technique and receiver noise (photon counting, speckle noise and background noise) (see “Appendix 1”).
More advanced and complex models [13, 17, 35, 38, 48, 49], can be used to simulate the propagation of the laser beam, for example, from laser energy distribution to atmospheric effects and beam interaction models, to receiver optics and received signal processing electronics models. However, this study is focusses on evaluating the processing speed of the proposed TIR approach when used with the LADAR simulator. Standard laser propagation models are used as this will preserve the generality and give a clear indication about performance under fundamental (standard) models.
3 Proposed Approach to Generating The Target Impulse Response
In order to generate a target impulse response for every laser burst, a laser beam footprint that illuminates the target surface \(I(H_{ls},V_{ls},R_{ls})\) must be created by using Eq. 4. The reflected power (\(P^{sample}_{i}\)) reaching the receiver from each sample in this laser footprint and the corresponding roundtrip time (i), are then calculated.
The sample reflected power is calculated from the LADAR range Eq. 8, while the roundtrip time is calculated by obtaining the intersection for this sample with the target’s surface. The reflected sample powers \(P^{sample}_{i}\) are then summed with the same time indices i to create the target impulse response (\(h_{tr}\)).
The barycentric coordinates [18] for this point, with respect to the triangle’s vertices, are then calculated to determine if it lies inside the triangle’s edges. In general, a point is inside (or on) the triangle if, and only if \(0\le w_{1}\le 1, 0\le w_{2} \le 1\), and \(w_{1}+w_{2} \le 1\).
 1.
The angular extant in terms of azimuth and elevation angular ranges for each triangle is calculated and stored. These calculation are required ones per scanning setup.
 2.
Laser ray vectors (right side of Fig. 3) are generated. These vectors depend on the LADAR viewing direction, laser footprint size and the number of laser footprint samples.
 3.
The triangles whose angular extents (calculated in step 1) lie within the laser beam illumination direction, are selected (the blue edges triangles in Fig. 3).
 4.
For every selected triangle, the laser ray vectors that lie in the field of that triangle are selected and the intersection points between each calculated using Eqs. 9, 10, and 11. The top right side of Fig. 3, shows the selected rays that lie in the field of the green edged triangle. It also shows the intersection points on the triangle plane (green & yellow points) and inside the triangle itself (green points).
 5.
If the laser ray vector lies in the field of more than one triangle and has intersection points with each, the point that has the shorter distance to the laser is selected and stored.

Azimuth angular range: This is computed by calculating the azimuth angle for each triangle’s vertices and comparing these angles with each other to find the minimum and the maximum values, these representing the azimuth angular range.

Elevation angular range: The method for calculating this range is similar to the method above except that the elevation angles for the triangle’s vertices do not always represent the range. Therefore, additional three edge angles (one per triangle edge) are calculated and added to the comparison.
4 LADAR Simulator
 1.
The required simulation parameters are defined; LADAR (viewing direction, field of view and scanning resolution), laser source (temporal and spatial domains), atmosphere, target, noise and receiver.
 2.For every laser pulse, the target impulse response \(h_{tr}\) is generated and convolved with the temporal laser pulse p(t) to calculate the temporal reflected power signal arriving at the detector \(P_{r}(t)\) as shown in the following equation.$$P_{r}(t)=h_{tr}(i)*p(t)$$(15)
 3.
The resultant power signals are converted to photoelectrons using Eq. 18. The background, photon counting, and speckle noise are then applied, if enabled by the user, using Eqs. 19 and 17 (see “Appendix 1”).
 4.The received electrical signals are then passed to the CFD peak detector to detect their peaks which are then used to calculate the roundtrip time intervals \(\Delta t_{tot}\) as shown in Fig. 1. As the laser pulses travel at the speed of light c (\(3\times 10^{8}\) m / s), the LADAR system calculates the ranges \(R_{c}\) for these pulses using the following equation [34]:$$R_{c}=\frac{\Delta t_{tot}}{2}\times c$$(16)
 5.
Finally the range values are assigned to the corresponding pixels on the LADAR image.
5 Testing Procedure and Evaluation Results
The performance of the LADAR simulator, in terms of processing time required to generate simulated 3D images, has been evaluated. This evaluation was achieved with the simulator using two target impulse response TIR generation approaches; conventional (normal) and proposed.
Since the time required to generate TIR depends on the number of rays’ vectors (that represent the laser beam samples) and on the number of triangles (that represent the scene or model surface), the simulator only tested changes in the effects of these parameters. The other scanning parameters were kept constant, their values presented in Table 1 in “Appendix 3”. This table also shows the specifications for the computer that is used to run this test. Changing the number of vectors is achieved by changing the spatial sampling factor (\(sf_{s}\), see Sect. 2.1.2), while changing the number of triangles is done by using different 3D models of different numbers of faces.
The testing procedure starts by scanning the 3D model using both approaches at a scanning resolution equal to 2500 pixels, with a laser beam (number of vectors equal to 768 by setting \(sf_{s}\) to 5), and calculates the required time to get the final image. The procedure then increases \(sf_{s}\) by 5 and rescans the model again until the \(sf_{s}\) reaches 50 this equivalent to 68403 vectors. Another 3D model comprised of more triangles than the previous model is then selected and the whole procedure is repeated again and so on, until 42 different 3D models are scanned. In order to guarantee that all models are fully scanned with the same resolution (2500 pixels), both the angular field of view and angular resolution are automatically adjusted according to the model dimensions.
In order to present these effects individually, some results have been selected from the original 3Dgraph and their slopes are also calculated (using the least square method) as shown in Fig. 11c, d. Figure 11c shows the effect of changing the number of triangles on execution time (and its slope) for specific vectors numbers (Vc. No.) while Fig. 11d shows the effect of changing the number of vectors on execution time (and its slope) for specific numbers of triangles (Tr. No.).
The results in Fig. 11b shows that the execution time for the proposed approach is much smaller than the normal approach. Figure 11c shows an increment in slopes for both approaches when the numbers of vectors increases from 768 to 68403. The differences in execution time when increasing the numbers of triangles, are not significant with the proposed approach. This caused the effect of models shapes clearly noticeable as a fluctuation in time (see Fig. 11c). Figure 11d also shows an increment in slopes for both approaches, but this time when the numbers of triangles increases from 2375 to 13980. In general, the average execution times for both normal and proposed approaches, are equal to \(6.1\times 10^{3}\) s and 17.6 s respectively.
6 Conclusion
An efficient LADAR simulator has been developed using a novel TIR generation approach, to simulate the direct detection time of flight LADAR systems. The simulator models each stage, from laser source to data generation, over a short execution time producing simulated LADAR images, under a wide variety of conditions. The proposed approach to generate TIR has been developed to produce responses identical to these generated from the conventional or standard approach, but by using less computational time. This has been achieved by mathematically deriving the required equations to calculate target angular ranges which, in turn, enables an evaluation of the intersection points that lie in the same angular range, instead of evaluating the whole intersection point (between every laser ray vector and all the scene’s triangles).
More than forty, 3D models were used to evaluate the simulator’s performance in terms of processing time with different laser beam samples. The evaluation was carried out with this simulator using two target impulse response TIR generation approaches, proposed and conventional (normal), where the latter was used to benchmark the results. A comparison of the results shows that the LADAR simulator with proposed approach is quicker than normal approach, especially when the 3D model consists of a large number of triangles, or when a large number of laser footprint samples are required.
The average processing speed for the simulator with the proposed approach was 345 times faster in comparison to the normal one. This improvement in performance enables the simulator to scan a large number of targets, at different scanning parameters and poses at high speed, and opens up the possibility for simulating LADAR beam propagation more accurately in a shorter time, by using a large number of laser footprint samples.
The simulation steps for the LADAR simulator and its GUI are illustrated with some results of scanned 3D models. These simulation results demonstrate the ability of the LADAR simulator to scan and produces LADAR images under different scanning parameters (noise type, scanning resolution and laser beam width).
Notes
Acknowledgements
I would like to express my sincere gratitude to those who have provided the financial support throughout the research, the Iraqi Ministry of Higher Education and Scientific Research. Also I would like to thank Dr. Toby Hall (Mathematical Sciences Department in University of Liverpool) for his help.
References
 1.Artist3d is a threedimensional objects download library. http://artist3d.com/.
 2.3dvia model library (2007). http://www.3dvia.com/.
 3.Accetta, J. S., & Shumaker, D. L. (1993). The infrared and electrooptical systems handbook. In C. S. Fox (Ed.), Active ElectroOptical Systems (Vol. 6). Infrared Information Analysis Center and SPIE Optical Engineering Press.Google Scholar
 4.AlTemeemy, A. A., & Spencer, J. W. (2011). Simulation of 3D ladar imaging system using fast target response generation approach. In Proceedings of SPIE—the international society for optical engineering (Vol. 8167). doi: 10.1117/12.902309.
 5.AlTemeemy, A., & Spencer, J. (2014). Laser radar invariant spatial chromatic image descriptor. Optical Engineering,. doi: 10.1117/1.OE.53.12.123109.Google Scholar
 6.AlTemeemy, A., & Spencer, J. (2015). Chromatic methodology for laser detection and ranging (ladar) image description. Optik, 126(23), 3894–3900. doi: 10.1016/j.ijleo.2015.07.182.CrossRefGoogle Scholar
 7.AlTemeemy, A., & Spencer, J. (2015). Invariant chromatic descriptor for ladar data processing. Machine Vision and Applications, 26(5), 649–660. doi: 10.1007/s0013801506750.CrossRefGoogle Scholar
 8.AlTemeemy, A. A., Spencer, J. W., & Ralph, J. F. (2010). Levy flights for improved ladar scanning. In 2010 IEEE international conference on imaging systems and techniques, IST (pp. 225–228). doi: 10.1109/IST.2010.5548519.
 9.Bachman, C. G. (1979). Laser radar systems and techniques. Norwood: Artech House.Google Scholar
 10.Barenz, J., Baumann, R., Imkenberg, F., Krogmann, D., & Tholl, H. D. (2004). All solid state imaging infrared/imaging ladar sensor system. Proceedings of SPIE—the international society for optical engineering (Vol. 5459, pp. 171–179).Google Scholar
 11.Bers, K., Schulz, K. R., & Armbruster, W. (2005). Laser radar system for obstacle avoidance. Proceedings of SPIE–the international society for optical engineering (Vol. 5958, pp. 1–10).Google Scholar
 12.Bostelman, R., Hong, T., & Madhavan, R. (2005). Obstacle detection using a timeofflight range camera for automated guided vehicle safety and navigation. Integrated ComputerAided Engineering, 12(3), 237–249.Google Scholar
 13.Budge, S., Leishman, B., & Pack, R. (2006). Simulation and modeling of return waveforms from a ladar beam footprint in usu ladarsim. In Proceedings of SPIE (pp. 263–268).Google Scholar
 14.Caldas, C.H., Haas, C.T., Liapi, K.A., & Teizer, J. (2006). Modeling job sites in real time to improve safety during equipment operation. In: Proceedings of SPIE—The International Society for Optical Engineering, vol. 6174 II, pp. 61743K (1–8).Google Scholar
 15.Coyac, A., Hespel, L., Riviere, N., & Briottet, X. (2015). Performance assessment of simulated 3D laser images using geigermode avalanche photodiode: Tests on simple synthetic scenarios. In Proceedings of SPIE—the international society for optical engineering (Vol. 9648). doi: 10.1117/12.2194304.
 16.De Haag, M. U., Venable, D., & Smearcheck, M.(2007). Use of 3d laser radar for navigation of unmanned aerial and ground vehicles in urban and indoor environments. In: Proceedings of SPIE—the international society for optical engineering (Vol. 6550, pp. 65500C (1–12)).Google Scholar
 17.Der, S., Redman, B., & Chellappa, R. (1997). Simulation of error in optical radar range measurements. Applied Optics, 36(27), 6869–6874.CrossRefGoogle Scholar
 18.Ericson, C. (Ed.). (2005). Real time collision detection. Burlington: Morgan Kaufmann.Google Scholar
 19.Ewald, A., & Willhoeft, V. (2000). Laser scanners for obstacle detection in automotive applications. In: Proceedings of IEEE intelligent vehicles symposium (pp. 682–687).Google Scholar
 20.Graham M., & Davies, A. M. D. (2009). Simulating foliagepenetrating ladar imagery. In: Proceedings of the SimTecT (pp. 682–687).Google Scholar
 21.Graham, M. (2009). Design of a foliage penetrating ladar simulation tool. General Document DSTOGD0577, The Defence Science and Technology Organisation , Australia’s Department of Defence, Australian.Google Scholar
 22.Hwang, S., Kim, S., & Lee, I. (2011). Efficient geometric computation for LADAR simulation. In Proceedings of SPIE—the international society for optical engineering (Vol. 8060). doi: 10.1117/12.883901.
 23.Jelalian, A. V. (1992). Laser radar systems. Norwood: Artech House.Google Scholar
 24.Jiménez, A. R., Ceres, R., & Seco, F. (2004). A laser rangefinder scanner system for precise maneouver and obstacle avoidance in maritime and inland navigation. In: Proceedings Elmar—international symposium electronics in marine (pp. 101–106).Google Scholar
 25.Jolivet, V., Fournier, J., Normandin, X., & Cariou, J. (2005). Feasibility of air target identification using laser radar vibrometry. In: Proceedings of SPIE—the international society for optical engineering (Vol. 5807, pp. 222–232).Google Scholar
 26.Kim, S., Hwang, S., Son, M., & Lee, I. (2011). Comprehensive highspeed simulation software for LADAR systems. In Proceedings of SPIE—the international society for optical engineering (Vol. 8181). doi: 10.1117/12.898401.
 27.Koechner, W. (2006). Solidstate laser engineering (springer series in optical sciences) (6th ed.). New York: Springer.Google Scholar
 28.Lengyel, E. (2004). Mathematics for 3D game programming and computer graphics (2nd ed.). Newton: Charles River Media.MATHGoogle Scholar
 29.Madhavan, R., & Hong, T. (2005). Robust detection and recognition of buildings in urban environments from ladar data. In: Proceedings—applied imagery pattern recognition workshop (pp. 39–44).Google Scholar
 30.Marino, R. M., Davis, W. R., Rich, G. C., McLaughlin, J. L., Lee, E. I., Stanley, B. M., et al. (2005). Highresolution 3d imaging laser radar flight test experiments. In: Proceedings of SPIE—the international society for optical engineering (Vol. 5791, pp. 138–151).Google Scholar
 31.Moon, H. C., Lee, H. C., & Kim, J. H. (2006). Obstacle detecting system of unmanned ground vehicle. In: 2006 SICEICASE international joint conference (pp. 1295–1299).Google Scholar
 32.Ng, T. C., Guzmán, J. I., & Tan, J. C. (2004). Development of a 3d ladar system for autonomous navigation. In: 2004 IEEE conference on robotics, automation and mechatronics (pp. 792–797).Google Scholar
 33.Redman, B. C., Stann, B., Lawler, W., Giza, M., Dammann, J., Ruff, W., Potter, W., Driggers, R. G., Garcia, J., Wilson, J., & Krapels, K.(2006). Chirped am ladar for antiship missile tracking and force protection 3d imaging: Update. In: Proceedings of SPIE–the international society for optical engineering (Vol. 6214, pp. 62140O (1–15)).Google Scholar
 34.Richmond, R. D., & Cain, S. C. (2010). Directdetection LADAR systems. Bellingham: SPIE Press.CrossRefGoogle Scholar
 35.Roth, B. D., & Fiorino, S. T. (2012). LADAR performance simulations with a high spectral resolution atmospheric transmittance and radiance model: LEEDR. In Proceedings of SPIE (Vol. 8379, pp. 83790O–83790O15). doi: 10.1117/12.918330.
 36.Saleh, B. E. A., & Teich, M. C. (2007). Fundamentals of photonics (2nd ed.). Hoboken: Wiley.Google Scholar
 37.Song, Z. (2004). 2d laser ray tracing for the simulation of laser perception. Tech. rep., College of Engineering, Utah State University.Google Scholar
 38.Steinvall, O., & Carlsson, T. (2001). Threedimensional laser radar modelling. Bellingham: SPIE Press.Google Scholar
 39.Trussell, C. W. (2001). 3d imaging for army applications. In: Proceedings of SPIE—the international society for optical engineering (Vol. 4377, pp. 126–131).Google Scholar
 40.Vadlamani, A. K., & De Haag, M. U. (2007). Aerial vehicle navigation over unknown terrain environments using inertial measurements and dual airborne laser scanners or flash ladar. In: Proceedings of SPIE—the international society for optical engineering (Vol. 6550, pp. 65500B (1–12)).Google Scholar
 41.Wallace, A. M., Buller, G. S., Sung, R. C. W., Harkins, R. D., McCarthy, A., HernandezMarin, S., et al. (2005). Multispectral laser detection and ranging for range profiling and surface characterization. Journal of Optics A: Pure and Applied Optics, 7(6), S438–S444.CrossRefGoogle Scholar
 42.Wallace, A. M., Sung, R. C. W., Buller, G. S., Harkins, R. D., Warburton, R. E., & Lamb, R. A. (2006). Detecting and characterising returns in a pulsed ladar system. IEE Proceedings: Vision, Image and Signal Processing, 153(2), 160–172.Google Scholar
 43.Welliver, M., Nichols, T., Gatt, P., Willis, C., & Bunte, S. (2005). Supercomputer based ladar signature simulator. In Proceedings of SPIE—the international society for optical engineering (Vol. 5791, pp. 354–365). doi: 10.1117/12.609693.
 44.Wijesoma, W. S., Kodagoda, K. R. S., & Balasuriya, A. P. (2004). Roadboundary detection and tracking using ladar sensing. IEEE Transactions on Robotics and Automation, 20(3), 456–464.CrossRefGoogle Scholar
 45.Xu, R., Shi, R., Wang, X., & Li, Z. (2014). Echo signal modeling of imaging LADAR target simulator. In Proceedings of SPIE—the international society for optical engineering (Vol. 9300). doi: 10.1117/12.2072700.
 46.Xu, R., Shi, R., Ye, J., Wang, X., & Li, Z. (2015). Research on key technologies of ladar echo signal simulator. In Proceedings of SPIE—the international society for optical engineering (vol. 9671). doi: 10.1117/12.2197990.
 47.Ye, C. (2007). Toward safe navigation in urban environments. In: Proceedings of SPIE—the international society for optical engineering (Vol. 6561, pp. 1–12).Google Scholar
 48.Youmans, D. G. (2012). Ladar imaging analytical approach using both outward and return path atmospheric turbulence phasescreens. In Proceedings of SPIE (Vol. 8379, pp. 83790F–83790F15). doi: 10.1117/12.919402.
 49.Youmans, D. G. (2014). Outward atmospheric scintillation effects and inward atmospheric scintillation effects comparisons for direct detection ladar applications. In Proceedings of SPIE (Vol. 9080, pp. 90800C–90800C15). doi: 10.1117/12.2050674.
Copyright information
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.