[工学]遗传算法在图像边缘检测中的应用.doc

内 容 提 要 边缘检测是图像处理的重要内容，是图像分析的基础。找到与图像中目标的实际边界线相对应的真实边缘，一直是图像处理领域的研究热点。对图像边缘检测算法的研究已经有几十年的历史，边缘检测算法逐渐走向成熟，但仍旧被列为图像处理的经典技术难题之一。这是因为传统的边缘检测算法基本上是基于梯度微分的，这些算法存在着不同程度的局限性。 遗传算法是借鉴生物界的自然选择和自然遗传机制的随机化有哪些信誉好的足球投注网站算法，是一种全局优化有哪些信誉好的足球投注网站算法，具有简单、通用、鲁棒性强、适于并行处理等特性。如何将遗传算法应用到图像的边缘检测中来，人们进行了大量研究，但是这些研究大部分以经典边缘检测算法为基础展开的。 本论文主要围绕遗传算法在边缘检测过程中的具体应用进行探讨，主要研究内容包括：首先对基于模板卷积的传统边缘检测算法进行了总结，进而研究了一种以遗传算法为优化工具对模板系数进行寻优的方法。结果表明该方法可以准确的在具有相同特征的图像上提取边缘。其次对遗传算法的基本策略及其收敛性进行了分析，重点研究了遗传算法的收敛性评价准则，探讨了一种能够有效克服早熟收敛的遗传算法。 关键词 图像处理；遗传算法；边缘检测 Application of Genetic Algorithms in Image Edge Detection Author: Tutor: Abstract Boundary detection plays an important role in image processing and analyzing. The task of finding the edges in an image that correspond to true physical boundaries remains a hotspot of image processing. Research on boundary detection has last for the past decades .During this period, the boundary detection algorithms had progressively developed. However, boundary detection algorithm still ranks as a difficult problem in the image processing. The traditional boundary detection algorithms are on the basis of gradient and have more or less limitations. Genetic Algorithm or GA is a kind of random search algorithm which uses natural selection and natural inheritance mechanism for reference. It is a kind of entirely optimizing search algorithm, which is provided with many characteristics, such as simpleness, currency, robustness, and fit for parallel processing. People did a lot of researches in how to use genetic algorithm to the image edge detection, but most of them are based on classical edge detection. In this thesis, research work focuses on how to use genetic algorithms to solve boundary detection problems. The main researches are as follows: First, the traditional boundary detection algorithms based on masks are summarized. Furthermore, a novel mask design method based on genetic algorithms is researched. The experimental results show that the given method can effectively extr