Yoann Altmann (personal webpage) and Julian Tachella (personal webpage) from Heriot-Watt University, UK, are presenting on Monday 2nd July their work on Bayesian methods for 3D reconstruction using single-photon Lidar.
The seminar will take place in the Salle de thèse, 2pm.
The development of photodetectors able to quantify light at a photonic level with enhanced timing capabilities has enabled the reconstruction of 3D scenes under extreme environments (long-range, underwater). In this talk, we will discuss how Bayesian methods can be used to extract information from highly uncertain data while providing cues about the quality of the information extracted. We will also see how such methods can be used within the acquisition process to produce more efficient imaging systems. To a first approximation, each single-wavelength LIDAR waveform consists of a main peak, whose position depends on the target distance and whose amplitude depends on the target reflectivity. When considering multiple wavelengths, it becomes possible to use spectral information in order to identify and quantify the main materials in the scene, in addition to the estimation of the LIDAR-based range profiles.
Due to the sparse and discrete nature of the data recorded (times of photon arrival), classical statistical methods relying on Gaussian noise assumptions usually fail in extreme scenarios (short acquisitions, high background counts) for which strong regularisation is required. By coupling structured/hierarchical Bayesian models with advanced MCMC methods or computationally less intensive optimisation methods, the results demonstrate the possibility to analyse the spectral profile of 3D scenes constructed from extremely sparse photon counts (1 per pixel and per band) and are extremely encouraging for long-range and fast hyperspectral imaging.