Search

This poster presentation was prepared for the Optical Architectures for Displays and Sensing in Augmented, Virtual, and Mixed Reality (AR, VR, MR) IV Conference at the SPIE AR | VR | MR 2023 Symposium., Scientific and Technological Research Council of Turkey (TÜBİTAK); 2247-A National Lead Researchers Program; European Union (EU); European Innovation Council (EIC); Horizon-EIC-2021-Transition-Challenges Program; JST FOREST Program Grant; JSPS KAKENHI; Royal Society Research Grants

International audience; Due to the limited number of photons, directly imaging planets requires long integration times. The wavefront must be stable on the same time scale which is often difficult in space due to thermal variations and vibrations. In this paper, we discuss the results of implementing a dark hole maintenance (DHM) algorithm (Pogorelyuk et. al. 2019)1 on the High-contrast imager for Complex Aperture Telescopes (HiCAT) at the Space Telescope Science Institute (STScI). The testbed contains a pair of deformable mirrors (DMs) and a lyot coronagraph. The algorithm uses an Extended Kalman Filter (EKF) and DM dithering to predict the drifting electric field in the dark hole along with Electric Field Conjugation to cancel out the drift. The DM dither introduces phase diversity which ensures the EKF converges to the correct value. The DHM algorithm maintains an initial contrast of 8.5 x 10-8 for 6 hrs in the presence of the DM actuator random walk drift with a standard deviation of 1:7 x 10-3 nm/s..

International audience; This article focuses on Group Target Tracking (GTT) to counter swarms of drones using Random Finite Sets (RFSs) and Random Matrix (RM) approaches. Tracking swarms of drones is analog to tracking extended targets that are characterized by their continuously evolving shape and composition. Extended target tracking for groups of targets finds various applications in the literature because detecting and tracking each individual target of a group is computationally demanding and unnecessary if the group itself can be modeled. Elliptic shapes offers a suitable representation for most groups, and their inference is quite inexpensive with the random matrix approach. Indeed, they are efficient when coupled with random finite sets based filters, which represents the current state of the art for Bayesian multi-target tracking. In this work, a practical implementation of a labeled Poisson/Multi-Bernoulli filter using random matrices for group target tracking is proposed. This study compares several random matrix prediction and update algorithms with and without random finite sets based filters on several dataset.

Cloud detection is a critical issue for satellite optical remote sensing, since potential errors in cloud masking can be translated directly into significant uncertainty in the retrieved downstream geophysical products. The problem is particularly challenging when only of a limited number of spectral bands is available, and thermal infrared bands are lacking. This is the case of Proba-V instrument, for which the European Space Agency (ESA) carried out a dedicated Round Robin exercise, aimed at intercomparing several cloud detection algorithms to better understand their advantages and drawbacks for various clouds and surface conditions, and to learn lessons on cloud detection in the VNIR and SWIR domain for land and coastal water remote sensing. The present contribution is aimed at a thorough quality assessment of the results of the cloud detection approach we proposed, based on Cumulative Discriminant Analysis. Such a statistical method relies on the empirical cumulative distribution function of the measured reflectance in clear and cloudy conditions to produce a decision rule. It can be adapted to the user's requirements in terms of preferred levels for both type I and type II errors. In order to obtain a fully automatic procedure, we choose as a training dataset a subset of the full Proba-V scenes for which a cloud mask is estimated by a consolidated algorithm (silver standard), that is from either SEVIRI, MODIS or both sensors. Within this training set, different subsets have been setup according to the different types of surface underlying scenes (water, vegetation, bare land, urban, and snow/ice). We present the analysis of the cloud classification errors for a range of such test scenes to yield important inferences on the efficiency and accuracy of the proposed methodology when applied to different types of surfaces.

International audience; This paper presents the analysis of the influence of de-noising algorithms from the point of view of the contrast transfer function (CTF). The study focuses on amplitude images from digitally recorded holograms of the USAF target. In order to assess the evaluations, a database is constituted with experimental images extracted from 20 holograms acquired with a random illumination. Then, using the 20 amplitude images, 20 other images are computed in order to get images with increasing SNR (signal to noise ratio). In order to study the influence of de-noising algorithms on the CTF, 16 ROIs (region of interest) are selected in the USAF target. Each ROI corresponds to a specific pattern at a given spatial resolution. For the evaluations, 34 denoising algorithms were chosen considering their efficiency in image processing and digital holography. We choose advanced methods as stationary wavelet transform based algorithm with Daubechies and symlets wavelet, curvelets, contourlets, BM3D algorithm (state of the art in the image processing) and NL-means algorithms; in addition, we consider classical methods such as Wiener, median and Gauss filtering, anisotropic filtering and Frost filter which was widely used in SAR (synthetic aperture radar) imaging. From the ROIs, the CTF is evaluated. Then, we provide ranking of de-noising algorithms by considering the measured CTF. The variation of the CTF versus the input SNR is analysed for each algorithm.

International audience; Plenoptic cameras provide single-shot 3D imaging capabilities, based on the acquisition of the Light-Field, which corresponds to a spatial and directional sampling of all the rays of a scene reaching a detector. Specific algorithms applied on raw Light-Field data allow for the reconstruction of an object at different depths of the scene.Two different plenoptic imaging geometries have been reported, associated with two reconstruction algorithms: the traditional or unfocused plenoptic camera, also known as plenoptic camera 1.0, and the focused plenoptic camera, also called plenoptic camera 2.0. Both systems use the same optical elements, but placed at different locations: a main lens, a microlens array and a detector. These plenoptic systems have been presented as independent. Here we show the continuity between them, by simply moving the position of an object. We also compare the two reconstruction methods. We theoretically show that the two algorithms are intrinsically based on the same principle and could be applied to any Light-Field data. However, the resulting images resolution and quality depend on the chosen algorithm.

International audience; This paper gives new insights on optical 3D imagery. In this paper we explore the advantages of laser imagery to form a three-dimensional image of the scene. 3D laser imaging can be used for three-dimensional medical imaging and surveillance because of ability to identify tumors or concealed objects. We consider the problem of 3D reconstruction based upon 2D angle-dependent laser images. The objective of this new 3D laser imaging is to provide users a complete 3D reconstruction of objects from available 2D data limited in number. The 2D laser data used in this paper come from simulations that are based on the calculation of the laser interactions with the different meshed objects of the scene of interest or from experimental 2D laser images. We show that combining the Radom transform on 2D laser images with the Maximum Intensity Projection can generate 3D views of the considered scene from which we can extract the 3D concealed object in real time. With different original numerical or experimental examples, we investigate the effects of the input contrasts. We show the robustness and the stability of the method. We have developed a new patented method of 3D laser imaging based on three-dimensional reflective tomographic reconstruction algorithms and an associated visualization method. In this paper we present the global 3D reconstruction and visualization procedures.

To the best of our knowledge, proposed is a novel variable depth of field smart imager design using intelligent laser targeting for high productivity multiple barcodes reading applications. System smartness comes via the use of an Electronically Controlled Variable Focal-Length Lens (ECVFL) to provide an agile pixel (and/or pixel set) within the laser transmitter and optical imaging receiver. The ECVFL in the receiver gives a flexible depth of field that allows clear image capture over a range of barcode locations. Imaging of a 660 nm wavelength laser line illuminated 95-bit one dimensional barcode is experimentally demonstrated via the smart imager for barcode target distances ranging from 10 cm to 54 cm. The smart system captured barcode images are evaluated using a proposed barcode reading algorithm. Experimental results after computer-based post-processing show a nine-fold increase in barcode target distance variation range (i.e., range variation increased from 2.5 cm to 24.5 cm) when compared to a conventional fixed lens imager. Applications for the smart imager include industrial multiple product tracking, marking, and inspection systems.

Object tracking is widely used in many applications such as intelligent surveillance, scene understanding, and behavior analysis. Graph-based semisupervised learning has been introduced to deal with specific tracking problems. However, existing algorithms follow- ing this idea solely focus on the pairwise relationship between samples and hence could decrease the classification accuracy for unlabeled samples. On the contrary, we regard tracking as a one- class classification issue and present a novel graph-based semi- supervised tracker. The proposed tracker uses linear neighborhood propagation, which aims to exploit the local information around each data point. Moreover, the manifold structure embedded in the whole sample set is discovered to allow the tracker to better model the target appearance, which is crucial to resisting the appearance variations of the object. Experiments on some public-domain sequen- ces show that the proposed tracker can exhibit reliable tracking performance in the presence of partial occlusions, complicated back- ground, and appearance changes, etc. © 2013 SPIE and IS&T (DOI