臺灣博碩士論文加值系統

本論文之主要目標為偵測影像的邊、線特徵及端點、角點、交點的位置。在邊線偵測方面，早期偵測法於靠近角點、交點附近的部份會斷裂；而本論文提出了區域灰度值調整的濾波方式，搭配方向通道及Teager能量組合來解決這個問題，並使偵測位置更加準確。在端點、角點、交點偵測方面，則根據方向通道的特性，提出對偶通道相乘法，可正確反應端點、角點、交點的能量分佈而得到其位置。

The main purpose of this thesis is to develop an image feature detection scheme based on local gray level modification. The extracted image features include edges, lines, terminals, corners and junctions. Most traditional edge and line detection methods failed to produce continuous contours near and at corners or junctions. This difficulty is resolved by performing a local gray level transformation to a given image before conducting the orientational filtering procedure. Followed by Teager energy computation and oriented energy peak detection, edges and lines are extracted with an appropriate threshold. On the other hand, terminals, corners and junctions are detected from the two-dimensional energy map that is generated by summing the energies of all the pairs of “opponent channels”. The energy of a pair of opponent channels is defined as the product of Teager energies from two orthogonal orientational channels. From experimental results, it has shown that the developed feature detection scheme produces continuous and correct edge and line features.

第一章緒論……………………………………………………… 1
1.1簡介 …………………………………………………… 1
1.2研究動機及流程 ……………………………………… 2
第二章文獻回顧與問題陳述…………………………………… 4
2.1邊線模型 ……………………………………………… 4
2.2傳統偵測邊線的方法 ………………………………… 5
第三章方向通道 ………………………………………………… 11
3.1高斯微分濾波器 ……………………………………… 13
3.2Teager 能量組合 ……………………………………… 15
3.3等方向性與非等方向性高斯濾波器 ………………… 20
3.3.1 等方向性高斯濾波器 ……………………………… 20
3.3.2非等方向性高斯濾波器 …………………………… 21
3.4邊線判斷準則 ………………………………………… 23
3.5偵測結果 ……………………………………………… 24
第四章正規化的方向通道 ……………………………………… 27
4.1建立於區域灰度值調整的濾波器 …………………… 28
4.2灰度值調整的方式 …………………………………… 29
4.3轉移函數的動作原理 ………………………………… 33
4.4方向通道輸出結果與傳統方式比較 ………………… 42
4.4.1輸入影像振幅之影響 ……………………………… 42
4.4.2 各方向通道響應之差異 …………………………… 48
4.4.3對調諧方向之影響 ………………………………… 48
4.4.4對非調諧方向之影響 ……………………………… 51
4.4.5對雜訊之影響 ……………………………………… 58
4.4.6門檻值的設定 ……………………………………… 59
4.5成功因素整理 …………………………………………59
第五章特徵偵測 ………………………………………………… 61
5.1邊、線偵測 ……………………………………………61
5.1.1三角形偵測結果 …………………………………… 61
5.1.2Ben圖偵測結果……………………………………… 63
5.1.3Teager能量組合法是否必要 …………………………65
5.1.4123比234次微分之使用時機…………………………66
5.1.5Lena影像偵測結果……………………………………68
5.1.6非等方向性濾波器對邊線偵測之影響………………69
5.2端點、角點、交點偵測 ………………………………71
5.2.1使用非等方向性濾波器………………………………73
5.2.2轉移函數的影響………………………………………73
5.2.3使用誤差函數偵測結果 …………………………… 74
5.2.4使用階梯函數偵測結果 …………………………… 75
5.2.5不使用轉移函數偵測結果比較………………………75
5.2.6Ben圖偵測結果 ………………………………………76
第六章整合多解析度之輸出 …………………………………… 78
6.1輸入信號寬度對濾波器標度之關係 ………………… 78
6.1.1最佳匹配濾波器 …………………………………… 79
6.1.2較大標度濾波器的彈跳問題 ……………………… 80
6.2搜尋法 ………………………………………………… 83
6.3自然影像的實驗結果 ………………………………… 85
第七章討論 ……………………………………………………… 88
7.1抑制法之可行性 ……………………………………… 88
7.2結合Canny之最大梯度方向法 ………………………… 88
7.3其他偵測法之結果比較 ……………………………… 91
7.4模擬人類視覺系統 …………………………………… 93
第八章結論 ……………………………………………………… 94

[1]J. Canny, “A Computational Approach to Edge Detection,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. PAMI-8, No. 6, pp.679-698, Nov., 1986.[2]C. Hsin, “Edge Extraction based on a Statistical Decision Theory,” The Second Symposium on Computer and Communication Technology, Taichung, Taiwan, pp. 104-116, 1996.[3]C.A. Rothwell, J.L. Mundy, W. Huffman and V.D. Nguyen, “Driving Vision by Topology,” Int’l Symp. Computer Vision, pp. 395-400, Coral Gables, Fla., Nov. 1995.[4]J. H. Elder and S. W. Zucker, “Local Scale Control for Edge Detection and Blur Estimation,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 20, No. 7, pp. 699-716, 1998.[5]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。[6]F. Mokhtarian and R. Suomela, “Robust Image Corner Detection through Curvature Scale Space,” IEEE Transactions on Analysis and Machine Intelligence, Vol. 20, No. 12, pp.1376-1381, Dec., 1998.[7]L. Ding, A. Goshtasby, “On the Canny edge detector,” Pattern Recognition, Vol. 34, pp.721-725, 2001.[8]F. Heitger, L. Rosenthaler, R. V. D. Heydt, E. Peterhans and O. Kubler, “Simulation of Neural Contour Mechanisms: from Simple to End-stopped Cells,” Vision Research, Vol. 32, No. 5, pp.963-981, 1992.[9]F. Heitger, R. V. D. Heydt, E. Peterhans, L. Rosenthaler and O. Kubler, “Simulation of Neural Contour Mechanisms: Representing Anomalous Contours,” Image and Vision Computing, Vol. 16, pp.407-421, 1998.[10]B. Robbins and R. Owens, “2D Feature Detection Via Local Energy,” Image and Vision Computing, Vol. 15, pp.353-368, 1997.[11]C. Hsin and Y. C. Chuang, “Low-level Image Representation based on Frames of Gaussian Scaling Functions and Gaussian Derivative Wavelets,” Proceeding of IPPR, pp.515-522, Taichung, Aug., 1997.[12]R. Hamila, J. Astola, F. A. Cheikh, M. Gabbouj and M. Renfors, “Teager Energy and the Ambiguity Function,” IEEE Transactions on Signal Processing, Vol. 47, No. 1, pp.260-262, Jan, 1999.[13]R.C. Gonzalez and R. E. Woods, “Digital Image Processing,” Book, Prentice Hall, 2002.[14]S. Ando, “Image Field Categorization and Edge/Corner Detection from Gradient Covariance,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 22, No. 2, pp.179-190, Feb, 2000.[15]R. P. Wurtz and T. Lourns, “Corner Detection in Color Images through a Multiscale Combination of End-stopped Cortical Cells,” Image and Vision Computing, Vol. 18, pp.531-541, 2000.