图像关联规则挖掘研究

Research of image association rules mining

介绍了图像关联规则的相关概念,描述了传统的双种群遗传算法的执行过程;针对采用固定染色体交叉概率和染色体变异概率容易出现早熟、收敛速度较慢等问题,设计出了能自适应调整的染色体交叉算子和变异算子。最后将改进后的双种群遗传算法成功地运用到Landsat卫星遥感图像,实现了图像关联规则的提取,为退耕还林决策提供了有力的依据。

Firstly, this paper discussed definitions of image association rules and associated terms, described the process of the traditional doule population genetic algorithm. Secondly, proposed an new scheme that the double population genetic algorithm used the self-adaptable crossover operator and mutation operator according to the shortage of early convergence and stagnation in the traditional double population genetic algorithm based on unconvertible rate of crossover operator and mutation op- erator. Lastly, successfully applied the improved double populations genetic algorithm to mine the image association rules from Landsat satellite remote sensing images. It is useful for decreasing farmland to increase forest.