Abstract
Monitoring environmental evolutions, one of the most crucial axes on which sustainable development is based, requires the knowledge of information on the observed geographical scenes. Due to the continuous technological developments in the remote sensing field, the data sources increase exponentially over time, and the information contained in satellite images becomes increasingly rich. These characteristics make the extraction, processing and resolution of such data a complicated task that varies according to the situation. It is, therefore, necessary to develop tools adapted to satellite images interpretation and analysis problems. In this context, the constraint satisfaction problem (CSP) seems to be one of the methods to solve these problems. Despite some challenges, the CSP approach has performed well and proven its value in various fields. The effectiveness of CSP is based on two key points: first, the definition of constraints, and second, the choice of resolution methods. Therefore, a synthesis document covering CSP resolution methods, both static and dynamic, becomes relevant and necessary.
This paper represents the first comprehensive review of CSP. We begin by listing CSP methods, detailing their principles, and presenting the corresponding algorithms. We then illustrate the execution process of CSP methods using examples applied in the remote sensing domain for each one. Following this, we present a complete comparative study arranged according to key characteristics, which is intended to guide researchers to help them select the most appropriate resolution method for any given context. Finally, we present a set of challenges and future directions designed to suggest and drive further research in this promising field.
•A comprehensive survey of resolution methods for constraint networks: static and dynamic algorithms•This paper represents the first comprehensive review of CSP resolution methods.•A guide of the most appropriate CSP resolution method for a given context•A set of challenges and future directions designed to suggest and drive future research in the field of CSP.