模糊邊緣檢測(cè)算法用于人工目標(biāo)的提取【源代碼+開題報(bào)告+畢業(yè)論文】


目 錄
摘 要
第一章 緒論 4
1.1 利用邊緣檢測(cè)進(jìn)行目標(biāo)提取的含義及其應(yīng)用領(lǐng)域 4
1.2 利用遙感影像進(jìn)行目標(biāo)提取的流程 4
1.3 模糊邊緣檢測(cè)算法用于人工目標(biāo)提取的構(gòu)想 7
1.4 畢業(yè)論文研究的主要內(nèi)容 7
第二章 利用模糊邊緣檢測(cè)算法提取人工目標(biāo)的基本理論 9
2.1 模糊邊緣檢測(cè)算法的基本原理 9
2.2 傳統(tǒng)邊緣檢測(cè)算法的缺陷 11
2.3 改進(jìn)模糊模糊邊緣檢測(cè)算法 11
第三章 其他相關(guān)的目標(biāo)提取算法及技術(shù)比較 13
3.1 基于改進(jìn)的模糊邊緣檢測(cè)算法的衛(wèi)星遙感影像邊緣的提取 13
3.2 基于改進(jìn)的模糊邊緣檢測(cè)算法與其他幾種常用算子檢測(cè)結(jié)果的比較 13
第四章 利用模糊邊緣檢測(cè)算法進(jìn)行人工目標(biāo)提取方法的研究與詳細(xì)設(shè)計(jì) 16
4.1 遙感影像可視化的研究及設(shè)計(jì) 16
4.2 模糊邊緣檢測(cè)算法的研究與設(shè)計(jì) 17
4.3 檢測(cè)結(jié)果影像二值化的研究及設(shè)計(jì) 18
4.4 邊緣檢測(cè)結(jié)果去噪和細(xì)化的研究和設(shè)計(jì) 19
4.5 直線段或曲線段的自動(dòng)檢測(cè)和擬合的研究 20
第五章 對(duì)房屋和立交橋等人工地物進(jìn)行目標(biāo)提取的實(shí)現(xiàn) 22
5.1 遙感影像可視化的實(shí)現(xiàn) 22
5.2 模糊邊緣檢測(cè)的實(shí)現(xiàn) 24
5.3 圖像二值化的實(shí)現(xiàn) 25
5.4 利用邊緣檢測(cè)提取目標(biāo)形狀特征即目標(biāo)提取的實(shí)現(xiàn) 25
第六章 結(jié) 論 26
參考文獻(xiàn) 28
致 謝 30

 
摘 要
遙感影像自動(dòng)提取人工地物不僅是攝影測(cè)量與遙感領(lǐng)域的一大難題也是計(jì)算機(jī)視覺與圖像理解領(lǐng)域研究的一個(gè)重點(diǎn)問題。因此，構(gòu)建基于遙感影像的地物目標(biāo)的自動(dòng)半自動(dòng)提取算法，對(duì)于遙感影像的判讀，分析和解譯具有重要科研價(jià)值和實(shí)際應(yīng)用價(jià)值。
本論文主要研究利用模糊邊緣檢測(cè)算法進(jìn)行人工目標(biāo)提取的設(shè)計(jì)思想和實(shí)現(xiàn)方式,并以研究所得理論為指導(dǎo)編寫出了包含模糊邊緣檢測(cè)等常用算法的用于遙感影像人工地物目標(biāo)提取的軟件Edge Detector1.0.0。
遙感影像上目標(biāo)的邊緣在影像上表現(xiàn)為灰度的不連續(xù)性，傳統(tǒng)的邊緣檢測(cè)算子主要對(duì)邊緣信號(hào)和噪聲信號(hào)不加區(qū)分，往往在圖像邊緣對(duì)比度較大的情況下才能獲取較好的邊緣提取效果。模糊邊緣檢測(cè)方法是Pal和King在1983年提出的一種將模糊理論應(yīng)用于圖像特征提取的邊緣檢測(cè)方法，已經(jīng)在模式識(shí)別和圖像處理中獲得了很好的應(yīng)用，充分利用了圖像所具有的不確定性往往是由模糊性引起的這一特性。
本論文首先簡要描述了經(jīng)典模糊邊緣檢測(cè)算法的基本原理，然后從分析其缺陷入手，提出改進(jìn)后的算法思想，介紹了詳細(xì)的具體實(shí)現(xiàn)步驟，并通過對(duì)二維影像邊緣輪廓的提取,驗(yàn)證本文的改進(jìn)算法與傳統(tǒng)邊緣檢測(cè)算法經(jīng)典模糊邊緣檢測(cè)算法相比，效果更好；同時(shí)也將這種邊緣檢測(cè)算法與其他常用的邊緣檢測(cè)算法的檢測(cè)效果進(jìn)行了比較，分析了不同算法的優(yōu)劣和適用范圍。
本論文主要研究內(nèi)容總結(jié)如下：
（1）熟悉目標(biāo)提取流程：總結(jié)并歸納常用的目標(biāo)提取方法及其流程，根據(jù)檢測(cè)目標(biāo)的不同，總結(jié)出典型人工目標(biāo)提取適用的相關(guān)方法，并設(shè)計(jì)其相應(yīng)流程；
（2）能用VC＋＋實(shí)現(xiàn)基于模糊邊緣檢測(cè)算法的人工目標(biāo)提取，學(xué)習(xí)并掌握該方法的基本原理，思考如何進(jìn)行算法改進(jìn)。在理論研究的指導(dǎo)下設(shè)計(jì)并實(shí)現(xiàn)相應(yīng)算法并采用該算法嘗試對(duì)城市典型人工目標(biāo)（如房屋和立交橋）的提取；
（3）利用設(shè)計(jì)好的模糊邊緣檢測(cè)算法實(shí)現(xiàn)實(shí)際應(yīng)用中的遙感影像的典型人工目標(biāo)提??；
（4）將模糊邊緣檢測(cè)算法與常規(guī)目標(biāo)提取算法進(jìn)行比較，尋找該方法的適用場(chǎng)合和各種算法的優(yōu)劣之處；

關(guān)鍵字：邊緣檢測(cè)，模糊邊緣檢測(cè)算法，隸屬函數(shù)，人工目標(biāo)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.