臺灣博碩士論文加值系統

影像中的角點又稱為關鍵點是影像中的局部重要特徵，在影像的
辨識，定位等領域中扮演著十分重要的角色。影像經過與中心像素運
算後並使用壓縮轉換來提升對比，後利用多尺度多通道的高斯二階微
分濾波器偵測出特徵響應，再進一步偵測出高曲率的地方，接著透過
本文提出的改善方法將邊緣響應有效的降低，既可以達到關鍵點的準
確特性之外，又可以去除不要的假關鍵點，提高關鍵點偵測的性能。
最後再利用『ROC 曲線』與『Recall and Precision 曲線』，來評價
與比較整體的偵測系統。

Corners, junctions, and terminals represent prominent local features in
images. They are named keypoints. Keypoint detection is a vital step in
many applications such as pattern recognition and image registration. The
purpose of this thesis is to develop a keypoint detector based on local
contrast intensity. Initially an input image is enhanced by a compressive
mapping curve and then is transformed to a line-type image by computing
absolute local contrast. Subsequently, the local contrast intensity image is
applied to the multi-scale and multi-orientation Gaussian second-order
derivative filters. The outputs of the filters are used to detect the high
curvature points. False keypoints which occur at linear edges or in noisy
texture areas are eliminated by an automatic threshold scheme. Finally, the
performance of the proposed method was evaluated by both the receiver
operating characteristic curve and the recall and precision curve. In addition,
it was compared with other methods.

致謝 .............................................................. I
摘要 ............................................................. II
ABSTRACT ........................................................ III
目錄 ............................................................. IV
第一章 緒論 ....................................................... 1
1.1 簡介 ...................................................... 1
1.2 研究動機 .................................................. 1
1.3 研究方法 .................................................. 2
1.4 研究目標 .................................................. 2
第二章 文獻回顧與問題 ............................................. 3
2.1 文獻回顧 .................................................. 3
2.1.1 利用邊緣的特性........................................... 3
2.1.2 直接利用影像的灰階度進行角點偵測......................... 7
2.2 問題整理 ................................................. 14
第三章 關鍵點特徵擷取 ............................................ 15
3.1 建立區域對比強度影像 ........................................ 15
3.2 建立多標度方向通道 .......................................... 20
3.2.1 方向通道................................................ 20
3.2.2 高斯2 階微分濾波器...................................... 20
3.2.3 高斯微分濾波器的特性.................................... 21
3.2.4 多標度通道............................................. 23
3.3 兩兩通道正交 ............................................. 25
3.4 壓縮形式 ................................................. 27
3.5 整合方法 ................................................. 35
3.5.1 匹配權重係數............................................ 35
3.5.2 多標度對偶整合與能量整合................................ 36
3.5.3 等方向濾波與非等方向濾波的融合.......................... 39
第四章 改善方法 .................................................. 44
4.1 抑制機制 ................................................. 44
4.1.1 拉普拉斯遮罩 ............................................. 44
4.1.2 使用拉普拉斯遞迴 ......................................... 47
4.2 刪除機制 ................................................. 53
4.2.1 使用曲率度方法........................................... 53
4.2.2 曲率度方法的改進 ........................................... 54
第五章 實驗結果與性能評價 ........................................ 59
5.1 實驗結果 .................................................... 59
5.2 性能評價 .................................................... 65
5.2.1 ROC..................................................... 65
5.2.2 Recall and Precision.................................... 66
5.2.3 ROC VS Recall and Precision............................ 67
5.3 關鍵點的準度 ................................................ 76
第六章 結論與展望 ................................................ 78
參考文獻 ......................................................... 79

【1】 A. Rosenfeld and E. Johnston ,Angle Detection on Digital
Curves ,IEEE Trans. Computer ,1973 , C- 22 : 875- 878.
【2】 H. Freeman and L. S. Davis ,Corner-finding Algorithm for
Chain-coded Curves ,IEEE Trans. Computer, 1977 , C- 26: 297-
303.
【3】 U Ramer,An Iterative Procedure for the Polygonal
Approximation of Plane Curves , Computer Graphics Image
Process, 1972, 16 ( 1 ) :244- 256.
【4】 T Pavlidis, S L Horowitz, Segmentation of Plane Curves , IEEE
Trans. on Computers, 1974, 23 : 860- 870.
【5】 H. P. Moravec ,Towards Automatic Visual Obstacle
Avoidance,InteHigence,Cambridge,UK,1977：584.
【6】 C. G.Harris and M. J.Stephens ,A Combined Comer and Edge
Detector,Proceedings of the 4th Alvey Vision Conference,
Manchester,England,1988.
【7】 D. G. Lowe, Distinctive Image Features from Scale-invariant
Keypoints ,International Journal of Computer Vision,Vol. 60, No.
2, pp.91-110, 2004.
【8】 D. G. Lowe, Local Feature View Clustering for 3d Object
Recognition ,IEEE Conference on Computer Vision and Pattern
Recognition, pp.682-688,Kauai, Hawaii, December, 2001.
【9】 Y. Ke and R. Sukthankar, PCA-SIFT: A More Distinctive
Representation for Localimage Descriptors, IEEE Conference
on Computer Vision and Pattern Recognition, Vol. 2,
pp.506-513,Washington, DC, USA, June, 2004.
【10】 S.Smith and M. Brady ,A New Approach to Low Level Image
Processing,IJCV, 1997, 23 ( 1) : 45 -78.
【11】 F. Mokhtarian and R. Suomela , Robust Image Corner
Detection Through Curvature Scale Space,IEEE Transactions
on Pattern Analysis and Machine Intelligence,1998,20(12)：
1376-1381.
【12】 X. C. He ,Curvature Scale Space Corner Detector with
Recognition,IEEE Proceedings of the 17th International
Conference on Pattern Recognition,Cambridge,UK,2004：
791-794.
【13】 G.Cong and S.D.Ma ,Corner Enhancement in Curvature
Scale Space,Pattern Recognition,1998,31(10)：1491-1501.
【14】 X.Zhang,Robust Image Corner Detection Based on
multi-scale Curvature Product,Journal of Image and
Graphics,2007,7(12)：1270-1275.
【15】 F.Arrebola and F Sandoval ,Corner Detection and Curve
Segmentation by Multi resolution Chain Code Linking,Pattern
Recognition,2005,38(10)：1596+1614.
【16】 X.Zhang,Corner Detection Based on Gradient Correlation
Matrices of Planar Curves ,Pattern Recognition,2010,43：
1207-1223.
【17】 張獻堂,影像特徵邊線之非線性偵測理論,逢甲大學電子工
程研究所碩士學位論文,2002.
【18】 邵澤民、朱劍英、王化明,基於SUSAN 算法的分層快速
角點偵測華南理工大學學報2006,34(7):65-68.
【19】 張浩鵬、王宗義,基於小波的多尺度SUSAN 角點偵測,哈
爾濱工程大學,光電工程學報2009,36(11):100-1056
【20】 陳世章,影像輪廓偵測,逢甲大學通訊工程學系研究所碩
士論文,2004.
【21】 楊敏宏,影像輪廓偵測與分割,逢甲大學通訊工程研究所
碩士學位論文,2005.
【22】 林胤谷,基於群組化邊緣與交點特徵的影像封閉輪廓擷取
方法,逢甲大學通訊工程研究所碩士學位論文,2007.
【23】 J.A. Noble,Images as Functions and Sets,Image and Vision
Computer,1992,10(1):19-29.
【24】 R.S.Lin and C.H.Chu, A Modified Morphological Corner
Detector ,Pattern Recognition Letters,1998,19(3/4):27-28.
【25】 R.Laganiere,A Morphological Operatorfor Corner
Detection. PatternRecognition,1998,31(11):1643-1652.
【26】 常興治,高立群,利用圓結構元素的形態學角點檢測,計算
機輔助設計與圖形學學報,2007,10(19):1342-1347.