Abstract
This paper presents a new Bayesian model and associated algorithm for depth and intensity profiling using full waveforms from time-correlated single-photon counting measurements in the limit of very low photon counts (i.e., typically less than 20 photons per pixel). The model represents each Lidar waveform as an unknown constant background level, which is combined in the presence of a target, to a known impulse response weighted by the target intensity and finally corrupted by Poisson noise. The joint target detection and depth imaging problem is expressed as a pixelwise model selection and estimation problem, which is solved using Bayesian inference. Prior knowledge about the problem is embedded in a hierarchical model that describes the dependence structure between the model parameters while accounting for their constraints. In particular, Markov random fields (MRFs) are used to model the joint distribution of the background levels and of the target presence labels, which are both expected to exhibit significant spatial correlations. An adaptive Markov chain Monte Carlo algorithm including reversible-jump updates is then proposed to compute the Bayesian estimates of interest. This algorithm is equipped with a stochastic optimization adaptation mechanism that automatically adjusts the parameters of the MRFs by maximum marginal likelihood estimation. Finally, the benefits of the proposed methodology are demonstrated through a series of experiments using real data.
Highlights
Time-of-flight laser detection and ranging (Lidar) based imaging systems are used to reconstruct 3-dimensional scenes in many applications, including automotive [1]–[4], environmental sciences [5], [6], architectural engineering and defence [7], [8] applications
This is typically the case for free-space depth profiling on targets at very long distances based on the time-correlated single-photon counting (TCSPC) technique [9], which negotiates the trade-offs between range/intensity estimation quality, data acquisition time and output laser power
For the measurements reported the optical path of the transceiver was configured to operate with a fiber-coupled illumination wavelength of 841 nm, and a silicon single-photon avalanche diode (SPAD) detector
Summary
Time-of-flight laser detection and ranging (Lidar) based imaging systems are used to reconstruct 3-dimensional scenes in many applications, including automotive [1]–[4], environmental sciences [5], [6], architectural engineering and defence [7], [8] applications. We propose an algorithm for applications where the flux of detected photons is small and for which classical depth imaging methods [9] usually provide unsatisfactory results in terms of range and intensity estimation This is typically the case for free-space depth profiling on targets at very long distances based on the time-correlated single-photon counting (TCSPC) technique [9], which negotiates the trade-offs between range/intensity estimation quality, data acquisition time and output laser power. Classical Bayesian estimators associated with the joint posterior cannot be computed due to the complexity of the model, in particular because the number of underlying parameters (number of targets) is unknown and potentially large To tackle this problem, a ReversibleJump Markov chain Monte Carlo (RJ-MCMC) [15], [16] method is used to generate samples according to this posterior by allowing moves between different parameter spaces.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
