進階搜尋


   電子論文尚未授權公開,紙本請查館藏目錄
(※如查詢不到或館藏狀況顯示「閉架不公開」,表示該本論文不在書庫,無法取用。)
系統識別號 U0026-2108201214195500
論文名稱(中文) 可利用傾斜角度結構元素的二值化型態學
論文名稱(英文) Binary Morphology with Leaning Structuring Elements
校院名稱 成功大學
系所名稱(中) 電機工程學系專班
系所名稱(英) Department of Electrical Engineering (on the job class)
學年度 100
學期 2
出版年 101
研究生(中文) 蘇昱彰
研究生(英文) Yu-Chang Su
學號 N27961203
學位類別 碩士
語文別 中文
論文頁數 72頁
口試委員 指導教授-陳中和
口試委員-賴源泰
口試委員-蘇文鈺
中文關鍵字 Mathematical Morphology  Binary  Spatially Variant  Leaning Structuring Elements  Convex 
英文關鍵字 Mathematical Morphology  Binary  Spatially Variant  Leaning Structuring Elements  Convex 
學科別分類
中文摘要 自數學型態學(Mathematical Morphology)被研究並發展以來,其被廣泛的應用在醫學、礦業、自動化處理等各個領域。隨著VLSI及FPGA製程技術的進步,不管其在於面積的縮小或運算速度上均日益提升,因此由硬體實現Mathematical Morphology運算,以期降低硬體面積使用率或降低運所需時脈週數(clock)等的演算法於近年陸續被提出。
Mathematical Morphology基礎的四個運算是擴張 (Dilation)、腐蝕 (Erosion)、斷開 (Opening)、閉合 (Closing)運算。而斷開、閉合運算是由擴張及腐蝕所組成,意即在實現Mathematical Morphology運算的架構各種研究裡,便是針對提出改善實現擴張及腐蝕這兩種運算的架構為目標。
一般的二值化(Binary)型態學運算大部分採用Delay-Line的架構,其優點在於結構元素形狀不會受到限制,以及在足夠的FIFO數量下,原圖輸入只要完成一次的掃瞄,便可以得到整張圖片的Mathematical Morphology運算結果,但缺點在於進行腐蝕或擴張運算的同時,此種架構佔用了大量的硬體資源,這在需要多次組合並重複擴張及腐蝕運算下,此種架構的模組無法進行大量的複製以提高平行度。
在Hugo Hedberg 發表的Binary Morphology With Spatially Variant Structuring Elements: Algorithm and Architecture [1]一文中,所提出的硬體架構,非常節省硬體資源的使用,並且其提出的面積可變異的結構元素(Spatially Variant Structuring Elements)架構,但其有效點連續距離累加的演算法也為其所能使用的結構元素形狀設下限制若要使用凸包(Convex)形狀結構元素的情況下,必須採行分解多次再合併結果的運算。
本片論文提出了改善上述論文的演算法及架構,在受限的結構元素圖形中,提出使用傾斜結構元素(Leaning Structuring Elements)進行運算時不需要分解結構元素進行多次處理的架構。在Mathematical Morphology運算當中,越多可用的結構元素形狀,代表其的應用範圍增加以及為了達到目標所需要的組合運算次數減少。本論文亦會與其他二值化型態學演算法所需的演算時脈週期及佔用資源進行比較。
英文摘要 Mathematical morphology (MM) is extensively applied in medical science, mining industry, automatic processing, and various fields since it has been developed. With the advancement of VLSI and FPGA manufacturing technology, chip downsizing and processing speed are improved gradually. As a result, algorithms concerning reducing hardware space utility rate or lowering processing clock cycle via hardware MM computing have been continuously proposed in recent year.
The four fundamental processing are dilation, erosion, opening, and closing processing. Yet the opening and closing are composed of dilation and erosion processing. In various studies about performing MM processing are equal to improving dilation and erosion processing.
Generally Binary morphology adopts Delay-Line architecture, in which SE shape is not confined served as its merit. With sufficient FIFO quantity, the MM processing result will be obtained by single original picture scanning. However, one drawback is that mass hardware resource is occupied under such processing architecture when performing dilation or erosion processing. In such an architecture that requires multiple combination and repetition of dilation and erosion processing, parallelism cannot be raised through huge amount duplicating.
A study, Binary Morphology With Spatially Variant Structuring Elements: Alogorithm and Architecture by Hugo Hedberg, Petr Dokladal, and Viktor Owall proposed a hardware architecture which save hardware resource. It also proposed Spatially Variant Structuring Element. Nevertheless, the algorithm of serially effect pixel accumulation was also confined by SE shape cannot be convex. As a result, while using convex SE, SE must be discomposed and precede multiple MM processing.
Based on the algorithms and structures of prior studies, the present study proposed that Leaning Structuring Elements structure can be adopted toward confined convex images. In MM processing, the more applicable SE shapes represent the increase of MM applications and the reduction of necessary combination processing. The present study will also compare the necessary processing clock and occupied resource with the other binary morphology.
論文目次 第1章 序論…………………………………………………………………12
1.1 Motivation………………………………………………………………………………………………12
1.2 Contribution…………………………………………………………………………………………………13
1.3 Organization of Thesis………………………………………………………………………13
第2章 Mathematical Morphology簡介與硬體實現的架構………14
2.1 Mathematical Morphology……………………………………………………………………14
2.1.1 Binary Dilatio………………………………………………………………………………………15
2.1.2 Binary Erosion………………………………………………………………………………………16
2.1.3 Binary Opening and Closing………………………………………………………17
2.2 Delay-Line Architecture……………………………………………………………………19
2.3 Binary Morphology with Spatially Variant Structuring Elements…………………22
2.3.1 Spatially Variant Structuring Elements………………………22
2.3.2 Structuring Elements………………………………………………………………………26
2.3.3 Hedberg’s Algorithm…………………………………………………………………………28
2.3.4 Handling the Borders………………………………………………………………………34
2.3.5 Hedberg’s Architecture…………………………………………………………………37
2.3.5.1 Update Stage………………………………………………………………………………………38
2.3.5.2 FF-chain…………………………………………………………………………………………………38
2.3.5.3 Compare Stage……………………………………………………………………………………38
2.3.5.4 Other Units…………………………………………………………………………………………39
2.3.6 Limitation and Extensions…………………………………………………………39
2.3.6.1 Limitation……………………………………………………………………………………………39
2.3.6.2 Extensions……………………………………………………………………………………41
2.3.7 Implementation Result and Performance…………………………43
2.3.8 Conclusion…………………………………………………………………………………………………44

第3章 可利用傾斜結構元素的二值化型態學演算法……………………47
3.1 Proposed Algorithm…………………………………………………………………………………47
3.1.1 Leaning Structuring Elements…………………………………………………47
3.1.2 Using Hedberg’s Algorithm…………………………………………………………48
3.1.3 Improved Algorithm……………………………………………………………………………49
3.2 Handling the Borders……………………………………………………………………………56
3.3 Modification of Architecture………………………………………………………59
第4章 實驗環境及實驗結果………………………………………………61
4.1 實驗環境…………………………………………………………………61
4.2 實驗結果…………………………………………………………………61
第5章 結論與未來展望……………………………………………………65
5.1 結論………………………………………………………………………65
5.2 未來展望…………………………………………………………………68
參考文獻………………………………………………………………………70
參考文獻 [1] Hugo Hedberg, Petr Dokladal, and Viktor Öwall, “ Binary Morphology With Spatially Variant Structuring Elements: Algorithm and Architecture ’’, IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 18, NO. 3, MARCH 2009.
[2] Computer Vision CITS4240,School of Computer Science & Software Engineering The University of Western Australia pp.2.
[3] Hugo Hedberg, Fredrik Kristensen, and Viktor Öwall, , “ Low-Complexity Binary Morphology Architectures With Flat Rectangular Structuring Elements’’, IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS—I: REGULAR PAPERS, VOL. 55, NO. 8, SEPTEMBER 2008.
[4] Frank Y.Shih, “IMAGE PROCESSING and MATHEMATICAL MORPHOLOGY Fundamentals and Applications’’ ,pp.19 - properties 10.
[5] Hugo Hedberg, Fredrik Kristensen, and Viktor Öwall, , “ Low-Complexity Binary Morphology Architectures With Flat Rectangular Structuring Elements’’, IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS—I: REGULAR PAPERS, VOL. 55, NO. 8, pp.2218, SEPTEMBER 2008.
[6] Hugo Hedberg, Petr Dokladal, and Viktor Öwall, “ Binary Morphology With Spatially Variant Structuring Elements: Algorithm and Architecture ’’, IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 18, NO. 3, pp.562, MARCH 2009.
[7] Hugo Hedberg, Petr Dokladal, and Viktor Öwall, “ Binary Morphology With Spatially Variant Structuring Elements: Algorithm and Architecture ’’, IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 18, NO. 3, pp.563, MARCH 2009.
[8] Hugo Hedberg, Petr Dokladal, and Viktor Öwall, “ Binary Morphology With Spatially Variant Structuring Elements: Algorithm and Architecture ’’, IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 18, NO. 3, pp.565, MARCH 2009.
[9] Hugo Hedberg, Petr Dokladal, and Viktor Öwall, “ Binary Morphology With Spatially Variant Structuring Elements: Algorithm and Architecture ’’, IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 18, NO. 3, pp.566, MARCH 2009.
[10] Hugo Hedberg, Petr Dokladal, and Viktor Öwall, “ Binary Morphology With Spatially Variant Structuring Elements: Algorithm and Architecture ’’, IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 18, NO. 3, pp.565, MARCH 2009.
[11] Hugo Hedberg, Petr Dokladal, and Viktor Öwall, “ Binary Morphology With Spatially Variant Structuring Elements: Algorithm and Architecture ’’, IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 18, NO. 3, pp.565, MARCH 2009.
[12] Hugo Hedberg, Petr Dokladal, and Viktor Öwall, “ Binary Morphology With Spatially Variant Structuring Elements: Algorithm and Architecture ’’, IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 18, NO. 3, pp.565, MARCH 2009.
[13] Hugo Hedberg, Petr Dokladal, and Viktor Öwall, “ Binary Morphology With Spatially Variant Structuring Elements: Algorithm and Architecture ’’, IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 18, NO. 3, pp.566, MARCH 2009.
[14] Hugo Hedberg, Petr Dokladal, and Viktor Öwall, “ Binary Morphology With Spatially Variant Structuring Elements: Algorithm and Architecture ’’, IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 18, NO. 3, pp.567, MARCH 2009.
[15] Hugo Hedberg, Petr Dokladal, and Viktor Öwall, “ Binary Morphology With Spatially Variant Structuring Elements: Algorithm and Architecture ’’, IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 18, NO. 3, pp.568, MARCH 2009.
[16] Hugo Hedberg, Petr Dokladal, and Viktor Öwall, “ Binary Morphology With Spatially Variant Structuring Elements: Algorithm and Architecture ’’, IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 18, NO. 3, pp.571, MARCH 2009.
[17] Hugo Hedberg, Petr Dokladal, and Viktor Öwall, “ Binary Morphology With Spatially Variant Structuring Elements: Algorithm and Architecture ’’, IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 18, NO. 3, pp.570, MARCH 2009.
[18] Hugo Hedberg, Petr Dokladal, and Viktor Öwall, “ Binary Morphology With Spatially Variant Structuring Elements: Algorithm and Architecture ’’, IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 18, NO. 3, pp.571, MARCH 2009.
論文全文使用權限
  • 同意授權校內瀏覽/列印電子全文服務,於2014-09-11起公開。


  • 如您有疑問,請聯絡圖書館
    聯絡電話:(06)2757575#65773
    聯絡E-mail:etds@email.ncku.edu.tw