本文主要研究一種結合影像拼接與對比強化技術的環景照建構方法。關於影像拼接的部份，我們主要以 Brown-Lowe 對於影像拼接的探討為基礎，對每張影像使用尺度不變特徵轉換演算法提取特徵，對於各張影像的特徵進行特徵匹配，並且使用隨機抽樣一致法，消除錯誤匹配的特徵點對，並篩選出合適的單應性矩陣用於拼接的座標轉換，最後再利用多頻段融合技術，使得在拼接銜接處的色彩看起來更為自然。在完成影像拼接後，我們接續一種由 Hsieh-Shao-Yang 所提出的對比強化的自適應變分模型，其中使用 k -means 演算法將影像劃分成 k 個部分，對於此自適應的變分模型通過交錯最小化演算法中的分裂布雷格曼法將模型拆解成三個子問題處理，並使用最簡易色彩平衡讓整體飽和度提升。最後我們提供一系列的影像拼接數值實驗驗證此方法對建構環景照的有效性。

In this thesis, we study a method for constructing panoramas that combines image-stitching and contrast enhancement techniques. Based on the works of Brown and Lowe for image stitching, we use the scale-invariant feature transform algorithm to extract features for each image, perform feature matching for each image feature, apply the random sampling consensus method to eliminate the mismatched feature point pairs, and construct the appropriate homography matrix for the coordinate transformation, and finally use the multi-band blending technology to make the color difference at the splicing joints look more naturally. After completing the image stitching, we employ an adaptive variational model proposed by Hsieh-Shao-Yang to enhance the contrast of the stitched image, where we use the k -means algorithm to divide the image into k parts. The split Bregman method is applied to split the problem into three sub-problems to solve the minimization problem associated with the adaptive variational model. We also use the simplest color balance technique to increase the overall saturation. Finally, we provide a series of image stitching numerical experiments to demonstrate the effectiveness of the proposed combined method.

[1] A. B. Petro, C. Sbert, and J.-M. Morel, Automatic correction of image intensity
non-uniformity by the simplest total variation model, Methods and
Applications of Analysis, 21 (2014), pp. 91-104.

[2] C. Gatta, A. Rizzi, and D. Marini, ACE: an automatic color equalization algorithm,
Proceedings of the First European Conference on Color in Graphics, Image,
and Vision (CGIV02), 2002, pp. 316-320.

[3] D. G. Lowe, Object recognition from local scale-invariant features, Proceedings
of the International Conference on Computer Vision, 2 (1999), pp. 1150-1157.

[4] D. G. Lowe, Distinctive image features from scale-invariant keypoints, International
Journal of Computer Vision, 60 (2004), pp. 91-110.

[5] J.-M. Morel, A. B. Petro, and C. Sbert, Screened Poisson equation for image
contrast enhancement, Image Processing On Line, 4 (2014), pp. 16-29.

[6] J. S. Beis and D. G. Lowe, Shape indexing using approximate nearestneighbour
search in high-dimensional spaces, Conference on Computer Vision
and Pattern Recognition, (1997), pp. 1000-1006.

[7] T. Lindeberg, Scale-space theory: a basic tool for analyzing structures at different
scales, Journal of Applied Statistics, 21 (1994), pp. 225-270.

[8] M. Brown and D. G. Lowe, Recognising panorama, Proceedings of the 9th International
Conference on Computer Vision, (2003), pp. 1218-1225.

[9] M. Brown and D. G. Lowe, Invariant features from interest point groups,
Proceedings of the British Machine Conference, (2002), pp. 23.1-23.10.

[10] N. Limare, J.-L. Lisani, J.-M. Morel, A. B. Petro, and C. Sbert, Simplest color
balance, Image Process. On Line, 1 (2011), pp. 297-315.

[11] T. Lindeberg and B. M. ter Haar Romeny, Linear scale-space I: basic theory,
In: B. M. ter Haar Romeny (eds), Geometry-Driven Diffusion in Computer Vision,
(1994), pp. 1-38. Computational Imaging and Vision, Vol 1, Springer,
Dordrecht, Netherlands.

[12] P.-W. Hsieh, P.-C. Shao, and S.-Y. Yang, Adaptive variational model for contrast
enhancement of low-light images, SIAM Journal on Imaging Sciences, 13
(2020), pp. 1-28.

[13] https://pl.m.wikipedia.org/wiki/Plik:Standard_
deviation_diagram_(decimal_comma).svg

[14] http://www.tjxzj.net/1805.html#lg=1&slide=1

[15] https://www.researchgate.net/figure/
Example-of-a-multi-scale-representation-for-Lena-image-constructed-by-fig8_308747165

[16] https://blog.csdn.net/u012418573/article/details/
96731384

[17] https://blog.csdn.net/qinghuaci666/article/details/
81877169