基于数字图像处理技术的缺陷检测算法的分析.doc

毕业设计（论文）题 目 基于数字图像处理技术的 缺陷检测算法研究 摘 要随着数字图像处理技术在工业领域的发展，基于数字图像处理技术的机件缺陷检测得到了广泛的应用。低成本、无损伤、准确，快速，使得机器视觉在缺陷检测领域具有重要的意义。本题专门研究基于数字图像处理的缺陷检测算法，并在工业组态软件Xavis中实现，以便更好的应用于实际。本次毕设在缺陷检测识别中的算法作了较深入的研究。关于机件的表面缺陷检测，本次毕设针首先对可用于缺陷检测的图像处理的各个基本算法，比如图像滤波，图像增强，图像形态学处理，图像分割，边缘检测等进行了研究。针对待检测的图像的配准问题，采取Matlab对于基于harris角点检测的图像配准、基于灰度信息的模板匹配等各种算法进行了仿真。最终在毕设中采用了三种效果比较不错的配准算法。针对待检测的缺陷图像，研究并提出了基于信息熵的缺陷检测算法、基于图像形态学处理的缺陷检测算法。并自己的创新性的提出了基于灰度信息的模板匹配缺陷检测算法并予以实现。取得了预想中的成果，检测的表面划痕效果清晰明显。针对算法的封装与联合调试，在Visual Studio++下基于OpenCV视觉库，把检测算法封装成动态链接库文件，然后供工业视觉组态软件Xavis调用。经测试，程序运行、容错能力良好。检测程序适用范围也比较广。此次毕设也我熟悉并掌握Visual Studio C++软件、Xavis软件和图像处理以及OpenCV的知识。同时培养了良好的编程能力，分析和解决问题的能力，为以后的科研夯实了基础。关 键 词：缺陷检测；信息熵；模板匹配；Xavis；OpenCVABSTRACTWith the development of the digital image processing technology in the industry fields, the detection technology about mechanical defects based on digital image processing has got widely used. Low-cost, no damage, accurate result and fast diagnosing, makes the machine vision plays an important part in the field of surface defects detection. This thesis research specially about the algorithms of defects detection based on the digital image processing technology, and it is achieved in the Xavis, which is one of the industry configuration software. So it can be put into use better. This thesis researches the application in the detecting and distinguishing about defects in depth.We studied many elemental algorithms of the image processing about the surface defects detection, such as image filtering, image enhancement, image morphology and segmentation, edge detection.To slove the problem of the image registtration, we put forward the defect detection algorithms based on harris corner and template matching in gray level, and get the simulation results with Matlab. Finally we pick out three better image registration algorithms.About the disabled image to b