|
|
|
Land Use/Land Cover
|
Reconstruction of long term land cover changes by a maximum likelihood interpolation method using genetic algorithm
Masahiko Nagai,Ryosuke Shibasaki,Huang Shaobo
Center for Spatial Information Science, University of Tokyo
Cw-503, Block C, 4-6- 1 Komaba
Meguro- ku, Tokyo 153-8505, Japan
TEL & FAX: +81- 3- 5452- 6417
E mail nagaim@iis.u - tokyo.ac.jp
Japan
Abstract
Even though long term land cover change is very important in various fields such as global
environmental studies, only fragmentary data has been available. The interpolation method is applied to
reconstruct long term land cover changes from fragmentary observational data and knowledge of the
changes. Genetic- Algorithm (GA) is used as interpola tion method. This method is very advantageous
when the density of observational data is low because it can create most probable spatio-temporal
distribution of class variables under the fragmentary observational data and behavioral models.
Introduction
- Introduction
It is very important to have an adequate knowledge of long term land cover change for understanding
what is happening in the present and may happen in the future. Human activities have modified the
natural environment significantly, while it has recently become clear that during the last centuries the
intensity and scale of these influences have increased very much. Although long term land cover change
is very important, only fragmentary data has been available. The maximum likelihood interpolation
method using genetic algorithm is applied to reconstruct long term land cover.
- Introduction of genetic algorithm(GA)
Genetic algorithm (GA) is the search algorithm that is based on the mechanisms of natural selection and
evolution of natural genetics. The approach combines survival of the fittest among string structures.
Genetic algorithm is computational simple and powerful in their search without restrictive assumptions
about search spaces. In a simple genetic algorithm , five basic aspects are considered ; the representation
or coding of the problem, the initialization of the population, the definition of the evaluation function, the
definition of genetic operators, and the determination of parameters.
- Optimization scheme for nominal variable interpolation
Most of natural properties change along a continuous scale. Spatial continuity and temporal continuity
give rationale for interpolating fragmentary observational data. There are many models now for
knowledge and rules governing spatio-tem poral patterns and behavior of geographic objects. They can
provide more robust and quantitative basis for interpolating observational data. Reliability of result
estimated from model simulation can be improved by combining reliable observation data. It is
reasonable to assume that spatio- temporal events or the voxel- field of nominal variables should
maximize likelihood under give observational data and behavioral models. Observational data and
behavioral models can be integrated in the process of maximizing the likelihood of spatio-temporal
events. Genetic algorithm is applied as a optimization scheme because searching for the most likely
spatio-temporal or voxel-field of nominal data is a typical combinatorial optimization problem.
|
|
|
|
|
|
|
