模糊边缘检测算法用于人工目标的提取【源代码+开题报告+毕业论文】.rar

模糊边缘检测算法用于人工目标的提取【源代码+开题报告+毕业论文】


目 录
摘 要
第一章 绪论 4
1.1 利用边缘检测进行目标提取的含义及其应用领域 4
1.2 利用遥感影像进行目标提取的流程 4
1.3 模糊边缘检测算法用于人工目标提取的构想 7
1.4 毕业论文研究的主要内容 7
第二章 利用模糊边缘检测算法提取人工目标的基本理论 9
2.1 模糊边缘检测算法的基本原理 9
2.2 传统边缘检测算法的缺陷 11
2.3 改进模糊模糊边缘检测算法 11
第三章 其他相关的目标提取算法及技术比较 13
3.1 基于改进的模糊边缘检测算法的卫星遥感影像边缘的提取 13
3.2 基于改进的模糊边缘检测算法与其他几种常用算子检测结果的比较 13
第四章 利用模糊边缘检测算法进行人工目标提取方法的研究与详细设计 16
4.1 遥感影像可视化的研究及设计 16
4.2 模糊边缘检测算法的研究与设计 17
4.3 检测结果影像二值化的研究及设计 18
4.4 边缘检测结果去噪和细化的研究和设计 19
4.5 直线段或曲线段的自动检测和拟合的研究 20
第五章 对房屋和立交桥等人工地物进行目标提取的实现 22
5.1 遥感影像可视化的实现 22
5.2 模糊边缘检测的实现 24
5.3 图像二值化的实现 25
5.4 利用边缘检测提取目标形状特征即目标提取的实现 25
第六章 结 论 26
参考文献 28
致 谢 30

 
摘 要
遥感影像自动提取人工地物不仅是摄影测量与遥感领域的一大难题也是计算机视觉与图像理解领域研究的一个重点问题。因此，构建基于遥感影像的地物目标的自动半自动提取算法，对于遥感影像的判读，分析和解译具有重要科研价值和实际应用价值。
本论文主要研究利用模糊边缘检测算法进行人工目标提取的设计思想和实现方式,并以研究所得理论为指导编写出了包含模糊边缘检测等常用算法的用于遥感影像人工地物目标提取的软件Edge Detector1.0.0。
遥感影像上目标的边缘在影像上表现为灰度的不连续性，传统的边缘检测算子主要对边缘信号和噪声信号不加区分，往往在图像边缘对比度较大的情况下才能获取较好的边缘提取效果。模糊边缘检测方法是Pal和King在1983年提出的一种将模糊理论应用于图像特征提取的边缘检测方法，已经在模式识别和图像处理中获得了很好的应用，充分利用了图像所具有的不确定性往往是由模糊性引起的这一特性。
本论文首先简要描述了经典模糊边缘检测算法的基本原理，然后从分析其缺陷入手，提出改进后的算法思想，介绍了详细的具体实现步骤，并通过对二维影像边缘轮廓的提取,验证本文的改进算法与传统边缘检测算法经典模糊边缘检测算法相比，效果更好；同时也将这种边缘检测算法与其他常用的边缘检测算法的检测效果进行了比较，分析了不同算法的优劣和适用范围。
本论文主要研究内容总结如下：
（1）熟悉目标提取流程：总结并归纳常用的目标提取方法及其流程，根据检测目标的不同，总结出典型人工目标提取适用的相关方法，并设计其相应流程；
（2）能用VC＋＋实现基于模糊边缘检测算法的人工目标提取，学习并掌握该方法的基本原理，思考如何进行算法改进。在理论研究的指导下设计并实现相应算法并采用该算法尝试对城市典型人工目标（如房屋和立交桥）的提取；
（3）利用设计好的模糊边缘检测算法实现实际应用中的遥感影像的典型人工目标提取；
（4）将模糊边缘检测算法与常规目标提取算法进行比较，寻找该方法的适用场合和各种算法的优劣之处；

关键字：边缘检测，模糊边缘检测算法，隶属函数，人工目标提取。

Abstract

To automatically extract remote sensing images is not only a major problem in remote sensing technology, but also a key research area in computer vision and image understanding fields. Therefore, building a algorithms based on remote sensing images to extract object automatically or semi-automatically has important value for analysis and interpretation of remote sensing images and practical application.
The major content of the paper is to study how to extract target using fuzzy edge testing algorithm and to give a detailed way can be applicable in practice. Having been led by the theory，I designed and implemented the programming Edge Detector1.0.0 , including fuzzy edge testing algorithm and other common algorithm, could be used in analyzing remote sensing image.
The edge of features in remote sensing images is showed discrete, the traditional edge detection algorithm is a key to the edge signal and noise signal without distinction, often in the context of the larger picture of contrast gradient can obtain better results from the edge. Fuzzy edge testing algorithm is a theory put forward by King and Pal in 1983 ,which can be used in detecting features edge.The algorithm has been put in practice in pattern recognition and image processing, making good use of the uncertainty often caused by the fuzziness of images.
The paper first briefly describes of classical edge fuzzy detection, then from the analysis of weakness, with proposed improvements, exists in the classical algorithm a more effective algorithm finally is devised including detailed process. The production of edge testing using the improved algorithm has higher quality than the ones dealt with classical algorithm. Besides, comparison has been made between edge fuzzy detection and other common algorithm (e.g. Canny, Sobel, Kirsch, etc).We can acquire advantage and disadvantage of the algorithm mentioned above from the result of these experiment.
The content of the paper can be briefed as follows:
(1) Be familiare with the flow of obtaining target: sum up common methods
in target extraction and design algorithm for particular features.
(2) Implement a program to extrat target using VC++ based on the fuzzy edge testing algorithm and think about how to improve it.
(3) Use the program have been implemented to extract man-made target in remote sensing images,taking a cloverleaf junction or some buildings for example.
(4) Compare the improved fuzzy edge testing algorithm with other common
Algorithm in order to find their advantage and advantage in different situation

Keywords: Edge testing, Fuzzy edge testing algorithm, Membership function, Man-made target extraction.