International audienceThis paper presents a method for scene flow estimation from a calibrated stereo image sequence. The scene flow contains the 3-D displacement field of scene points, so that the 2-D optical flow can be seen as a projection of the scene flow onto the images. We propose to recover the scene flow by coupling the optical flow estimation in both cameras with dense stereo matching between the images, thus reducing the number of unknowns per image point. The use of a variational framework allows us to properly handle discontinuities in the observed surfaces and in the 3-D displacement field. Moreover our approach handles occlusions both for the optical flow and the stereo. We obtain a partial differential equations system coupling both the optical flow and the stereo, which is numerically solved using an original multi-resolution algorithm. Whereas previous variational methods were estimating the 3-D reconstruction at time $t$ and the scene flow separately, our method jointly estimates both in a single optimization. We present numerical results on synthetic data with ground truth information, and we also compare the accuracy of the scene flow projected in one camera with a state-of-the-art single-camera optical flow computation method. Results are also presented on a real stereo sequence with large motion and stereo discontinuities. Source code and sample data are available for the evaluation of the algorithm
The Céré-la-Ronde underground gas storage reservoir in the Paris Basin, a test site to study and enhance reservoir seismic monitoring, is a water-bearing sandstone reservoir in a faulted anticline structure. Seismic data acquired so far have generated a qualitative interpretation of the location of the gas bubble by studying fluid saturation (Meunier, 1998). However, between 1994 and 1997, two sonic logs showed subtle differences in V P not explained solely by saturation variations. Changes in pore pressure and stresses also influence reservoir elastic properties. Hence, we used geomechanical modeling to evaluate quantitatively how exploiting the gas reservoir impacts seismic measurements.Our method begins by computing, in a reservoir simulator, pore pressure and saturations. Pore pressure is a key input in the geomechanical modeling that produces mean effective stresses. These and the saturations are used to update seismic velocities in accordance with rock physics theory. In the final step, the introduction of a wavelet allows seismic modeling and the study of seismic attributes.Modeling. The reservoir simulation uses a 3-D finite volume code. The model covers 336 ǂ 228 km 2 . The key level of the reservoir model is split into three reservoir layers. R1 is used for gas storage. R2 and R3 are separated by shaly layers. A 2-D section was extracted from the 3-D model for the geomechanical modeling and overburden and underburden layers were added because the model has to cover the whole geologic column from the surface to a depth of 1500 m ( Figure 1).Geomechanical properties of each layer are extracted from core measurements (when available), sonic logs, or using characteristics of analogs. Initially, vertical stresses are determined by rock densities and horizontal stresses by an estimated stress ratio. The initial pore pressure is constant. The load for the geomechanical modeling is determined from cycling variations of the pore pressure in the reservoir (computed by the reservoir simulation). Our modeling has no lateral displacement at external boundaries. Figure 2 shows the results of the geomechanical modeling (i.e., pore pressure and mean total stresses variations) at well A for different production times. Mean effective stress (σeff) is defined as difference between mean total stress and pressure.Rock physics models are used to evaluate the impact of mean effective stresses on effective bulk modulus. We use contact models based on Hertz-Mindlin contact theory (Mindlin, 1949). In this technique, two identical spherical grains of radius R are deformed by normal and tangential forces. The radius of the contact area is a function of σeff, meaning effective shear and bulk moduli are also linked to mean effective stress. Using both moduli, V P and V S may be computed. Hertz-Mindlin theory assumes that velocity varies with σeff raised to the 1/6 th power. Some laboratory measurements on samples gave a smaller exponent: 0.09 for V P and 0.13 for V S . If the initial velocity (V 0 ) is known, the new velocity (V 1 )...
Abstract. After 2003's summer heat wave, Electricité deFrance created a global plan called "heat wave-dryness". In this context, the present study tries to estimate high river temperatures for the next decades, taking into account climatic and anthropogenic evolutions. To do it, a specific methodology based on Extreme Value Theory (EVT) is applied. In particular, a trend analysis of water temperature data is done and included in EVT used. The studied river temperatures consist of mean daily temperatures for 27 years measured near the French power plants (between 1977 and 2003), with four series for the Rhône river, four for the Loire river and a few for other rivers. There are also three series of mean daily temperatures computed by a numerical model. For each series, we have applied statistical extreme value modelling. Because of thermal inertia, the Generalized Extreme Value (GEV) distribution is corrected by the medium cluster length, which represents thermal inertia of water during extremely hot events. The µ and σ parameters of the GEV distributions are taken as polynomial or continuous piecewise linear functions of time. The best functions for µ and σ parameters are chosen using Akaike criterion based on likelihood and some physical checking. For all series, the trend is positive for µ and not significant for σ , over the last 27 years. However, we cannot assign this evolution only to the climatic change for the Rhône river because the river temperature is the resultant of several causes: hydraulic or atmospheric, natural or related to the human activity. For the other rivers, the trend for µ could be assigned to the climatic change more clearly. Furthermore, the sample is too short to provide reliable return levels estimations for return periods exceeding thirty years. Still, quantitative return levels could be compared with physical models for example.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.