This work presents a processing technique for enhancing images collected by an underwater modulated pulse laser imaging system. Laser-based sensors offer high-resolution and high-accuracy ranging in the underwater environment. However, these capabilities can be degraded in turbid waters due to scattering. This work presents experimental results demonstrating an image processing technique that reduces the effects of both backscatter and forward scatter. Without the use of gating, filtering, or a priori information, the processing technique can generate useful imagery to 6.9 attenuation lengths in a controlled laboratory environment.
Blue-green laser systems are being developed for optical imaging and ranging in the underwater environment. The imaging application requires high range resolution to distinguish between multiple targets in the scene or between multiple target features, while the ranging application benefits from measurements with high range accuracy. The group at the Naval Air Warfare Center Aircraft Division (NAWCAD) in Patuxent River, MD has been investigating the merging of wideband radar modulation schemes with a pulsed laser system for underwater imaging and ranging applications. For the imaging application, the narrow peak produced by pulse compression at the receiver offers enhanced range resolution relative to traditional short pulse approaches. For ranging, the selection of modulation frequency bands approaching 1GHz provides backscatter and forward scatter suppression and enhanced range accuracy. Both passband and baseband digital processing have been applied to data collected in laboratory water tank experiments. The results have shown that the choice of processing scheme has a significant impact on optimizing the performance of modulated pulse laser systems for either imaging or ranging applications. These different processing schemes will be discussed, and results showing the effect of the processing schemes for imaging and ranging will be presented.
Proc. SPIE. 10186, Ocean Sensing and Monitoring IX
KEYWORDS: Signal to noise ratio, Principal component analysis, Modulation, Backscatter, Scattering, LIDAR, Sensors, Photons, Signal processing, Environmental sensing, Statistical signal processing, Ranging, Absorption
This work presents a new statistical signal processing approach to reduce the effects of forward scatter on range accuracy for an underwater modulated pulse lidar. Lidar sensors offer the potential for high-resolution, high-accuracy ranging in the underwater environment. For the modulated pulse lidar rangefinder, performance is limited in turbid waters primarily due to forward scatter, which causes decreased range resolution and accuracy. This work presents simulated and experimental results demonstrating the ability of statistical signal processing to reduce range error for systems operating in these turbid conditions. Experimental results demonstrated 60% reduction in range error compared to a baseline approach.
Adaptive filtering and channel estimation techniques are applied to laser based ranging systems that utilize wide-band intensity modulation to measure the range and reflectivity of underwater objects. The proposed method aims to iteratively learn the frequency dependent characteristics of the underwater environment using a frequency domain adaptive filter, which results in an estimate for the channels optical impulse response. This work presents the application of the frequency domain adaptive filter to simulated and experimental data, and shows it is possible to iteratively learn the underwater optical channel impulse response while using Hybrid Lidar/Radar techniques.
This work demonstrates a new statistical approach towards backscatter “clutter” rejection for continuous-wave underwater lidar systems: independent component analysis. Independent component analysis is a statistical signal processing technique which can separate a return of interest from clutter in a statistical domain. After highlighting the statistical processing concepts, we demonstrate that underwater lidar target and backscatter returns have very different distributions, facilitating their separation in a statistical domain. Example profiles are provided showing the results of this separation, and ranging experiment results are presented. In the ranging experiment, performance is compared to a more conventional frequency-domain filtering approach. Target tracking is maintained to 14.5 attenuation lengths in the laboratory test tank environment, a 2.5 attenuation length improvement over the baseline.
The performance of a frequency-modulated continuous-wave (FMCW) hybrid lidar-radar system will be presented in the context of an underwater optical ranging application. In adapting this technique from the radar community, a laser is intensity-modulated with a linear frequency ramp. A custom wideband laser source modulated by a new wideband digital synthesizer board is used to transmit an 800 MHz wide chirp into the underwater channel. The transmitted signal is mixed with a reference copy to obtain a “beat” signal representing the distance to the desired object. The expected form of the return signal is derived for turbid waters, a highly scattering environment, indicating that FMCW can detect both the desired object and the volumetric center of the backscatter “clutter” signal. This result is verified using both laboratory experiments and a realistic simulation model of the underwater optical channel. Ranging performance is explored as a function of both object position and water turbidity. Experimental and simulated results are in good agreement and performance out to ten attenuation lengths is reported, equivalent to 100 meters in open ocean or 5 meters in a turbid harbor condition.
In this paper simulation and experimental results are presented for two hybrid lidar-radar modulation techniques for underwater laser ranging. Both approaches use a combination of multi-frequency and single frequency modulation with the goal of simultaneously providing good range accuracy, unambiguous range, and backscatter suppression. The first approach uses a combination of dual and single frequency modulation. The performance is explored as a function of increasing average frequency while keeping the difference frequency of the dual tones constant. The second approach uses a combination of a stepped multi-tone modulation called frequency domain reflectometry (FDR) and single frequency modulation. The FDR technique is shown to allow simultaneous detection of the range of both the volumetric center of the backscattered “clutter” signal and the desired object. Experimental and simulated results are in good agreement for both techniques and performance out to ten attenuations lengths is reported.