THE DECISION OF THE OPTIMAL PARAMETERS IN MARKOV RANDOM FIELDS OF IMAGES BY GENETIC ALGORITHM

This paper introduces the principle of genetic algorithm and the basic method of solving Markov random field parameters.Focusing on the shortcomings in present methods,a new method based on genetic algorithms is proposed to solve the parameters in the Markov random field.The detailed procedure is discussed.On the basis of the parameters solved by genetic algorithms,some experiments on classification of aerial images are given.Experimental results show that the proposed method is effective and the classification results are satisfactory.

The Integration of Cooperation Model and Genetic Algorithm

In the photogrammetry,some researchers have applied genetic algorithms in aerial image texture classification and reducing hyper_spectrum remote sensing data.Genetic algorithm can rapidly find the solutions which are close to the optimal solution.But it is not easy to find the optimal solution.In order to solve the problem,a cooperative evolution idea integrating genetic algorithm and ant colony algorithm is presented in this paper.On the basis of the advantages of ant colony algorithm,this paper proposes the method integrating genetic algorithms and ant colony algorithm to overcome the drawback of genetic algorithms.Moreover,the paper takes designing texture classification masks of aerial images as an example to illustrate the integration theory and procedures.

Research on the Perceptual Law of Artificial Ants

Beginning with the analysis of the behavior of natural ants, this paper illuminates the principle and method that, by adopting image texture energy as pheromone and finding their way on the track of the pheromone, artificial ants have the ability to identify and remember through similar measurement of pheromone. Based on the quantity of experiments, this paper analyzes some factors that influence the ability of artificial ants and draws some conclusions about the law of ant perception.

基于图像信息熵的图像分类

引用信息论中信息熵可以区分不同信息源包含不同信息量的思想,解决图像分类的问题。在图像分类中为了节省计算时间,将每幅100×100像元图像化算为以5×5像元构成的区域,作为大像元的图像,大像元中25个像元特征值的均值作为一个大像元的特征值,本文的图像分类是在大像元图像上进行的。首先计算出每幅大像元图像的信息熵,按照已知信息将图像分为5类,在每一类别中以图像总数的1/4的原则确定该类图像的取样数目,由每一类别中每幅大像元图幅的信息熵,计算各类别大像元图像取样的信息熵均值。在这个基础上,选择3类图像作为一个组合,计算待检验图像的信息熵Hl,分别与3个取样信息熵的均值HCPa、HCPb、HCPc之间的绝对值Δ1、Δ2、Δ3,其中Δi的最小者所属的类别,便是待检验图像的类别。通过实验证明本文提出的大像元信息熵,用于图像分类是可行的、有潜力的。

基于图像信息熵的高光谱图像分类

首先将一幅100像元×100像元的高光谱图像划分为以5像元×5像元构成的区域,作为大像元的图像,大像元中25个像元特征值的均值作为一个大像元的特征值。每组图像有175幅图像,实验中将他们分成四段,发现第三、四段的信息熵较小(其数值仅为二或三位数的数量级),仅用第一、二段的数据进行分类。选用图像信息熵≥1 500的图幅,进行该类图像信息熵均值的计算。农田、山体、居民地、水体4类不同地物的信息熵的均值和4种地物中某个地物的属性,利用其信息熵值与4种不同地物信息熵的均值比较,差值最小者的属性,即为待定地物的属性。

适用于训练样本选择的斜交因子模型研究

训练样本的质量直接影响训练阶段的训练质量(或效果),进而在一定程度上影响测试阶段的分类精度。训练样本的代表性和典型性则反映出训练样本质量的一个重要方面。对于当前非常流行的深度学习模型研究,如何尽可能地减少训练样本的数量,一方面成为一个非常“棘手”的问题,另一方面从实际应用的角度来看,这也上升为一个经济或成本方面的问题。提出了一种适用于训练样本选择的斜交因子模型方法,该方法松弛了Q型因子分析和对应分析对于公因子之间独立的假设条件,并在斜交参考解的基础上提出一种适合训练样本选择的近似求解斜交旋转的方法。实验结果表明,所提方法是可行、有效的。与基于正交因子模型的方法相比,它可以更好地描述或逼近现实的真实情况,可以选择出更合理、更具有代表性的典型训练样本,并且还可以取得满意的分类精度。适用于训练样本选择的斜交因子模型方法优于基于正交因子模型的训练样本的选择方法,被选择的训练样本分布相对更分散、更合理,而且总的分类精度平均提高3%左右。