ECCO 2024
Abstract Submission

71. Satellite Image Mosaic Selection Problem: Current state and research directions

Invited abstract in session TC-1: Algorithms, stream Algorithms.

Thursday, 11:30 - 13:00
Room: L226

Authors (first author is the speaker)

1. Jedrzej Musial
Institute of Computing Science, Poznan University of Technology
2. Jacek Blazewicz
Institute of Computing Science, Poznan University of Technology

Abstract

Satellite imaging solutions are widely used to study and monitor various regions of the Earth. However, a single satellite image can cover only a limited area. In cases where a larger area of interest is being examined, several images must be combined to create one larger image, called a mosaic.
As the number of available satellite images has increased significantly, selecting the optimal combination of images to build a mosaic has become increasingly difficult. The number of photos covering one location can reach hundreds. This is even more difficult if the user is interested in optimizing several parameters. Users must manually select the images they want to include on the cover without a computational approach. Providing a feasible solution (or selection of solutions) from which users can choose a cover is crucial to saving money and time, given that high-resolution satellite imagery is expensive.

The Satellite Image Mosaic Selection (SIMS) problem is a new optimization problem rapidly gaining attention due to its short-term practical application (NP-hard). In doing so, it is important to recognize that a problem can be defined in many ways or with many requirements and constraints. First of all, we can talk about a single-objective problem (mainly taking into account the cost) and a multi-objective problem (here there may be many goals). For all variants, various problem-solving techniques are used, such as sophisticated versions of branch-and-bound solutions, constraint programming, mixed-integer linear programming, Pareto local search algorithms, various solvers, etc.

SIMS problem modeling is expected to evolve rapidly in the near future, with many very practical variations. The most promising seems to be the development of a very accurate cloud cover model that will be able to almost perfectly reproduce the distribution of clouds in the image, as well as other emerging artifacts. You should also remember at least such features (which may be crucial) of the profession such as resolution, angle of incidence, time of taking a photo, temporal consistency of a series of photos, color level, contrast, saturation, and many others.

Research is partially supported by the Luxembourg National Research Fund (FNR) & the National Centre for Research and Development (NCBR) under the SERENITY Project (ref. C22/IS/17395419; POLLUX-XI/15/Serenity/2023).

Keywords

Status: accepted

Back to the list of papers