图像目标提取算法及其硬件实现--毕业论文设计.doc

河北大学工商学院 本科生毕业论文（设计） PAGE 32 图像目标提取算法及其硬件实现 摘　　要 边缘是图像最基本的特征，因此，边缘检测是图像处理中非常重要的内容。本文首先介绍了几种经典的目前被广泛使用的图像边缘检测算子，比如：罗伯特（Roberts）边缘算子、Prewitt边缘算子和索贝尔（Sobel）边缘算子等。这些算法的核心思想是假设边缘点对应于原始图像灰度级梯度的局部极值点。然后在 Matlab环境下对这些算子进行编程实现。但是，有些图像包含有噪声，这些算法对噪声非常敏感，常常会把噪声当作边缘点检测出来，而真正的边缘由于噪声的干扰也很有可能被漏检。其次，文章介绍了图像分割技术，其核心思想是：为了辨识和分析目标，需要将有关区域分离提取出来，并对目标进一步加以利用。在此基础上，利用这些算子对图像进行了边缘检测，总结出了各种边缘检测算法的优缺点，针对不同的图像找到相对比较有效的边缘检测算法，进而大幅度地减少了数据量，保留了图像重要的结构属性。本文还介绍了如何用硬件对图像进行边缘检测，并与用Matlab得到的边缘检测仿真结果进行了对比，证明了用硬件实现图像边缘检测的可行性。 关键词：边缘检测；图像分割；目标提取 Image Object Extraction Algorithm and Hardware Implementation ABSTRACT Edge is the most basic features of the image,therefore,，edge detection is very important in image processing content. This paper describes several classical and widely used in image edge detection operator，for example，Robert edge operator，Prewitt and Sobel edge operator，etc.The core idea of these algorithms is the assumption that the edge points corresponding to the original image gray level gradient of the local extreme point. Then in the Matlab environment of these operators for programming.However, some images contain noise,These algorithms are very sensitive to noise, the noise will often be detected as edge points, while the real edge due to noise is also likely to be missed. Secondly, the article introduces the image segmentation technique, its core idea is: to identify and analyze the goals, the need to extract the isolated region, and the goal of further use. On this basis, the use of these operators on the image edge detection, summed up the advantages and disadvantages of various edge detection algorithms for different images found relatively more effective edge detection algorithm, thus significantly reducing the amount of data, important to retain the structural properties of the image. This article also describes how to use the hardware of the image edge detection, and with the edge detection using Matlab simulation results obtained were compared to that of a hardware implement