棉花图像检索中的分块权重优化

Block Weight Optimization of Cotton Image Retrieval

针对棉花图像检索中颜色直方图方法中的各图像分块权重难以设定的难题,提出一种基于粒子群算法的多分块权重自动优化方法,它将图像各分块权重设定转化为智能求优过程,使用粒子群算法求取较佳分块权重组合。结合棉花数据集的实验结果表明,与均等权重方法以及用户自定义权重方法相比,该方法能够取得较为优化的权重组合,提高棉花图像的检索性能。

Aiming at the problem that it is difficult to set image block weight in cotton image retrieval with color histogram, this paper proposes multi-block weight automatic optimization method based on Particle Swarm Optimization（PSO）.It converts the block weight setting into intelligent optimization process,and then PSO is used to gain optimal block weight combination.Experimental results combined with cotton data set show that the method ca gain optimized weight combination and improves cotton image retrieval property,compared with equal weighting method and user-defined weighting method.