Search

We consider the problem of removing and replacing clouds in satellite image sequences, which has a wide range of applications in remote sensing. Our approach first detects and removes the cloud-contaminated part of the image sequences. It then recovers the missing scenes from the clean parts using the proposed "TECROMAC" (TEmporally Contiguous RObust MAtrix Completion) objective. The objective function balances temporal smoothness with a low rank solution while staying close to the original observations. The matrix whose the rows are pixels and columnsare days corresponding to the image, has low-rank because the pixels reflect land-types such as vegetation, roads and lakes and there are relatively few variations as a result. We provide efficient optimization algorithms for TECROMAC, so we can exploit images containing millions of pixels. Empirical results on real satellite image sequences, as well as simulated data, demonstrate that our approach is able to recover underlying images from heavily cloud-contaminated observations., To Appear In Conference on Computer Vision and Pattern Recognition (CVPR 2016)

This paper proposes and explores the properties of a solution algorithm designed to solve network simulation problems to optimality. Emphasis is put on the system being divisible into components, as this underlying assumption motivates the algorithm in its entirety in that rather simple relations between the system components are modeled as explicit structural constraints, while the significantly more complex relations within each component are approximated based on the underlying simulator data. The Real-Time petroleum Production Optimization (RTPO) problem is used as an example of such a network simulation problem that can benefit from using the solution approach outlined. The paper also includes the mathematical formulation of the RTPO, followed by a description of the algorithm and a comprehensive study of the algorithm's properties on a real RTPO application. Finally, the results are compared to a reference case and conclusions are reached about the method.

The recent success of shale-gas production relies on drilling of long horizontal wells and stimulation with multistage hydraulic fracturing. This practice normally leads to an initial peak production with a subsequent rate decline, followed by low and erratic production rates caused by water accumulation in the wells. Shale-gas recovery requires a large number of wells in order to maintain a sustainable total gas supply. To reduce the surface area disturbances caused by this extensive drilling and to share available surface infrastructure, the use of multi-well pads is a key driver in shale-gas developments. Furthermore, the inherent rate decline of shale-gas wells, the water accumulation in them and the large number of wells, leads to severe operational challenges for well operators. The fact that shut-ins may be used as a means to prevent liquid loading and boost late-life production rates from shale-gas wells, suggests scheduling of shut-ins to perform maintenance and clean-up of the wells, and to track a target rate for the multi-well pad. In this paper we propose an optimization scheme for shale-gas multi-well pads to schedule shut-ins and to track a target rate. The optimization problem is formulated as a discrete time mixed integer linear program (MILP) with binary variables dening at which times the well is either shut-in or producing. A reservoir proxy model and a well model for each well is designed and tuned against a realistic multi-fractured reservoir model. We demonstrate the benets and the potential of the proposed methodology through a one-month production planning problem for an eight-well shale-gas pad.