臺灣博碩士論文加值系統

防手震技術是為了減少或是避免因為手震所產生的模糊影像。此種技術能讓使用者更容易用相機照出好的照片。目前手機幾乎是人手一台，使用手機相機的人數眾多，使用者會在各種情況下利用手機相機拍照，而手震常常是造成照相品質不良的主因之一，所以這項技術已經成為目前手機相機需要具備的重要功能。
在這篇論文裡面，我們提出從四張影像合成出一張沒有手震影像的系統。利用降低曝光時間並且連續照出四張照片的方式，將四張照片合成，重建出一張亮度正常且為銳利的影像。由於此方法夠快且夠簡單因此能實做在一般的手機相機中。

Image stabilization aims to reduce the amount of image blur due to human handshake or even preventing the chances of a blur. This technique can allow people to take good pictures by camera phone easily. Because almost everyone owns a mobile phone today, many people use camera phone under various conditions. Handshake is one of the most important reasons to produce blurred image, so this technique has played an essential role in today’s camera phone design.

In this thesis, we present a digital image stabilization algorithm for camera phone based on an image composition technique using four source images. The method is to take four images with lower exposure time and then combine these four pictures. This method can reduce the amount of image noise and compose a sharper image from four source images. This method is simple and fast enough to be used by general camera phone.

目 錄
口試委員會審定書……………………………………………………………… i
誌謝………………………………………………………………………………. ii
中文摘要………………………………………………………………………… iii
英文摘要…………………………………………………………………………. iv
Chapter 1 Introduction……………………………………………………………. 1
1.1 Causes of Blurred Images………………………………………………… 2
1.2 Exposure Time and Focal Length……………………………………….. 3
1.3 Feature Detector…………………………………………………………. 7
Chapter 2 Optical Image Stabilization Systems…………………………………… 8
2.1 Common Approaches of Optical Image Stabilization…….……………… 9
2.2 Optical Image Stabilization by Moving Lens…..……………………….. 11
2.3 Optical Image Stabilization by Moving Sensor..……………………….. 12
2.4 Optical Image Stabilization of Vari-Angle Prism………………………. 14
Chapter 3 Digital Image Stabilization System……………………………….. 15
3.1 Digital Image Stabilization Using High ISO Speed…………………… 15
3.2 J. F. Chen’s Algorithm….............……………………………………… 16
3.3 Y. M. Lin’s Algorithm..…………………………………………………. 17
3.4 Some Product of Digital Image Stabilization on Camera Phone..………. 19
Chapter 4 Image Stabilization for Camera Phone……………………………….. 21
4.1 Pre-Blur………………….………………………………………………. 21
4.2 Feature Detection……………….……………………………………….. 22
4.3 Feature Matching..………………………………………………………. 24
4.4 Image Composition..……………………………………………………. 25
4.5 Flow Chart………..……..………………………………………………. 27
Chapter 5 Experimental Results, Conclusion and Future Work……………….. 28
5.1 Experimental Results……………………………………………………… 28
5.2 Conclusion……………………………………………………………… 40
5.3 Future work……………………………………………………………… 40
Reference ………………………………………………………………….…… 40




























圖 錄
Figure 1.1 Three types of blurred images………..………………………………… 2
Figure 1.2 The picture of moving water during different Exposure time..……….. 3
Figure 1.3 The relation between hand-shake and shutter speed….………………. 4
Figure 1.4 The relationship between number of sharp images and shutter speed.… 5
Figure 1.5 The relationship between view angle and focal length……………….. 5
Figure 1.6 Different focal lengths for the same hand-shake..……………………. 7
Figure 1.7 The result of corner detector…………………………………………… 8
Figure 2.1 Roll, Pitch, and Yaw, from NASA…………………….……………… 10
Figure 2.2 Image stabilization system of Nikon Vibration Reduction..………….. 10
Figure 2.3 High level block diagram of optical image stabilization….………….. 11
Figure 2.4 High level block diagram of optical image stabilization..……………. 14
Figure 2.5 Canon Optical Image Stabilizer………………….…….…………….. 12
Figure 2.6 Konica Minolta Anti-Shake System…….…..……………………….. 13
Figure 2.7 Konica Moving CCD structure..……………………………………… 13
Figure 2.8 Konica Minolta Moving CCD structure……………………………… 13
Figure 2.9 Canon Vari-Angle Prism…….……………………………………….. 14
Figure 3.1 Olympus digital image stabilization.………………………………… 15
Figure 3.2 J. F. Chen’s flow chart……………..……………………………….. 16
Figure 3.3 Y. M. Lin’s flow chart………………………………………………. 18
Figure 3.4 NTT DoCoMo FOMA N902i………………………………………… 19
Figure 3.5 Illustration of Shake-Correction…………………………………….. 20
Figure 3.6 N902i test images.…………………………………………………… 20
Figure 3.7 Noisy images to compare shake correction mechanism……………… 20
Figure 4.1 Images before and after blurring……………………………………… 22
Figure 4.2 Image with block artifact……..…………………………………….. 22
Figure 4.3 Cluster of feature points………………………………………………. 23
Figure 4.4 Image after filtering non-maximum feature points...………………… 24
Figure 4.5 Match pairs between two images…………………..……………….. 25
Figure 4.6 Images with and without only combining low frequency area……… 27
Figure 4.7 Our flow chart………………………………………………………… 27
Figure 5.1 Result image composed from four source images (Sony F828, 8 Mega-pixel, (a) exposure time: 1/2000 second, aperture: 2.8)..….……………………………. 28
Figure 5.2 Result image composed from four source images (Sony F828, 8 Mega-pixel, (a) exposure time: 1/40 second, (c) exposure time: 1/10 second)........................... 30
Figure 5.3 Result image composed from four source images (Sony F828, 8 Mega-pixel, (a) exposure time: 1/125 second (c) exposure time: 1/30 second)………………… 31
Figure 5.4 Result image composed from four source images (Sony F828, 8 Mega-pixel, (a) exposure time: 1/80 second, (c) exposure time: 1/20sec).……………………... 32
Figure 5.5 Result image composed from four source images (Sony F-828, 8 Mega-pixel, (a) exposure time: 1/160 second, (d) exposure time: 1/40 second).… 32
Figure 5.6 Result image composed from four source images (Sony F-828, 8 Mega-pixel, (a) exposure time: 1/160 second, (c) exposure time: 1/4 second).….. 34
Figure 5.7 Cropped images of Figure 5.6………………..……………………….. 36
Figure 5.8 Result image composed from four source images (Sony F-828, 8 Mega-pixel, (a) exposure time: 1/250 second, (c) exposure time: 1/60 second)…. 36

Figure 5.9 Cropped images of Figure 5.8………………..……………. 38
Figure 5.10 Result image composed from four source images (Sony F-828, 8 Mega-pixel, (a) exposure time: 1/160 second). …………………………………… 39

[1] A. Censi, A. Fusiello, and V. Roberto, “Image Stabilization by Features Tracking,” Proceedings of IEEE International Conference on Image Analysis and Processing, Venice, Italy, pp. 665-667, 1999.
[2] B. F. Robert, “The RANSAC (Random Sample Consensus) Algorithm,” http://homepages.inf.ed.ac.uk/rbf/CVonline/LOCAL_COPIES/FISHER/RANSAC/, 2002.
[3] Canon, “What is Optical Image Stabilizer,” http://www.canon.com/bctv/faq/optis.html, 2007.
[4] Canon, “What is Vari-Angle Prism,” http://www.canon.com/bctv/faq/vari.html, 2007.
[5] C. Harris and M. Stephens, “A Combined Corner and Edge Detector,” Proceedings of Alvey Vision Conference, Manchester, England, pp. 147-151, 1998.
[6] D. G. Lowe, “Distinctive Image Features from Scale-Invariant Keypoints,” International Journal of Computer Vision, Vol. 60, No. 2, pp. 91-110, 2004.
[7] G, Will, “Really Tiny Image Stabilization Systems for Camera Phones,” http://reviews.cnet.com/4531-10921_7-6531018.html, 2006.
[8] InvenSense, “Image Stabilization Technology,” http://www.invensense.com/company/technology.html, 2006.
[9] J. F. Chen, “Image Stabilization with Best Shot Selector and Super Resolution Reconstruction,” Master Thesis, Department of Computer Science and Information Engineering, National Taiwan University, 2005.
[10] K. Rockwell, “Why IS and VR Matter,” http://www.kenrockwell.com/tech/image-stabilization.htm, 2006.
[11] K. Willey, “Wide Angle and Telephoto Lenses,” http://www.kevinwilley.com/l3_topic03.htm, 2006.
[12] Konica Minolta, “Image Field,” http://konicaminolta.com/about/research/core_technology/picture/antiblur.html, 2007.
[13] M. Tico, S. Alenius, and M. Vehvilainen, “Method of Motion Estimation for Image Stabilization,” Proceedings of IEEE International Conference on Acoustics, Speech, and Signal Processing, Toulouse, France, Vol. 2, No. 2, pp. 277-280, 2006.
[14] NEC, “N902i-NECワイワイもばいる-カメラ,” http://www.n-keitai.com/n902i/cmr.html, 2005.
[15] Nikon, “VR (Vibration Reduction) System,” http://www.nikon.co.jp/main/eng/portfolio/about/technology/nikon_technology/vr_e/index.htm, 2005.
[16] Olympus, “Digital Image Stabilization,” http://www.olympusamerica.com/cpg_section/dismode.asp.
[17] Panasonic, “MEGA Optical Image Stabilization,” http://www2.panasonic.com/webapp/wcs/stores/servlet/MegaOISExplained?storeId=15001&catalogId=13401&langId=-1&catGroupId=24999, 2006.
[18] R. C. Gonzalez, R. E. Woods, and S. L. Eddins, Digital Image Processing Using MATLAB, Prentice-Hall, Upper Saddle River, New Jersey, 2004.
[19] S. David, N. Steven, and G. Danel, “Image Stabilization Technology Overview,” http://www.invensense.com/shared/pdf/ImageStabilizationWhitepaper_051606.pdf, 2007.
[20] S. Mallat, A Wavelet Tour of Signal Processing, Academic Press, San Diego, 1990.
[21] Wikipedia, “Focus,” http://en.wikipedia.org/wiki/Focus_(optics), 2007.
[22] Wikipedia, “Motion Blur,” http://en.wikipedia.org/wiki/Motion_blur, 2007.
[23] Wikipedia, “Shutter Speed,” http://en.wikipedia.org/wiki/Shutter_speed, 2007.
[24] Wikipedia, “Corner Detection,” http://en.wikipedia.org/wiki/Corner_detection, 2007.
[25] Wikipedia, “Feature Detection,” http://en.wikipedia.org/wiki/Feature_detection, 2007.
[26] Wikipedia, “Image Stabilization,” http://en.wikipedia.org/wiki/Image_stabilization, 2007.
[27] Wikipedia, “Singular Value Decomposition,” http://en.wikipedia.org/wiki/Singular_value_decomposition, 2007.
[28] Wikipedia, “Sobel,” http://en.wikipedia.org/wiki/Sobel, 2007.
[29] Y. M. Lin, “Image Stabilization System on a Camera Module with Image Composition,” Master Thesis, Department of Computer Science and Information Engineering, National Taiwan University, 2006.