灰度图像特征提取法研究毕业论文.doc

摘 要 图像特征提取一直是图像研究和计算机识别中一个值得探讨的问题，在计算机科学、医疗辅助诊断、军事、工业测量等各种领域都正在发挥着越来越重要的作用。尤其是医疗检索系统中，如何准确定位和提取关键特征往往是首要解决的问题之一，是提高医疗仪器识别率的关键。在医疗检索系统中准确的定位和提取图像的边缘特征，能够使医生快速准确的判别出病因。本文针对边缘特征提取算法和灰度直方图特征提取算法进行了特征提取试验与分析。 边缘特征提取，本文通过采取Robets算法，Sobel算法，Canny算法，小波变换算法对一幅图像做了不同的边缘特征提取。结果表明，本文采取的Canny边缘特征提取算子法明显优于其他边缘特征提取算法，且能够最为接近的描述灰度图像的边缘特征信息。 灰度直方图特征，对于边缘算子法提取出的灰度边缘特征图像，通过人眼观测有时可能会观测不到明显的区别。本文通过分析其灰度灰度直方图特征，提取其中的灰度特征量（灰度均值（Mean）、方差（Variance）、熵ntropy）、能量（Energy）），根据特征量的变化情况分析图像边缘提取效果的差异。 本算法编程采用C#语言，在Visual studio2010平台下实现，通过各种算法对图像特征提取的描述实现其不同的提取效果。结果显示C#语言能更好的表示各种不同算法在图像处理中的编程。 关键词：灰度图像；特征提取；边缘特征；灰度直方图；特征量 Abstract Research and computer image feature extraction has been a problem that worth exploring in visual identification, In computer science, medical diagnosis, military, industry measurement and so on various fields are playing an increasingly important role. especially in medical retrieval system, how t accurately locate and extract the key features are often the first to solve one of the problems, Is the key to improve medical instrument recognition rate. In medical retrieval system accurately locate and extract the image edge features, Can make the doctor quickly accurately identify the cause. Based on edge feature extraction algorithm and the algorithm for calculating grayscale histogram feature extraction feature extraction experiment and analyzed. The edge feature extraction, Algorithm in this paper, by adopting Robets, Sobel algorithm, the Canny algorithm, the wavelet transform algorithm to an image made of different edge feature extraction. Results show that the adopted Canny edge operator of feature extraction method is superior to other edge feature extraction algorithms, And can describe the most close to the edge of gray image feature information. Gray level histogram feature, for gray level edge operator method to extract the edge character images, by the human eye observation may sometimes observed no significant difference between the two. In this article, throu