近年來隨著科技的飛速發展，現今社會已是充斥著人工智慧與智慧型手機的時代，因此我們提出一種基於人工智慧的室內即時定位系統，目的是為了讓使用者藉由智慧型手機即可得知使用者在室內的位置訊息，也可降低架設與維護成本的一種高效率定位方式。現有且技術成熟的室內定位，基本上是藉由無線訊號(如藍芽等)作為定位方式，而這種方法除了設備與維護成本較高以外，還須考量無線訊號裝置的覆蓋率的問題，若覆蓋率不佳則有可能在某些地方因收不到訊號而無法在室內定位。本篇論文與以往的室內定位方式不同，是以不須依靠藍芽、WIFI等裝置，且完全依靠電腦視覺為主的室內定位論文。本論文模型是藉由現今已存在且研究成熟的物件辨識為基礎並延伸其功能的一種具有深度估計與角度估計的模型，進而定位出使用者於室內的位置訊息，可以方便使用者的操作，以博物館導覽為例，使用者可直接拿起智慧型手機並由相機掃描博物館內的任一物件，即可得知使用者於博物館內的位置，也可做延伸應用，如博物館導覽的路徑規劃等。

With the rapid development of technology in recent years, today's society is full of artificial intelligence and smartphones. Therefore, we propose an artificial intelligence-based indoor real-time positioning system, in order to allow users to know the user's indoor location information through smartphones, and to reduce the cost of installation and maintenance of an efficient positioning method.In addition to the higher cost of equipment and maintenance, the coverage of wireless signal devices must also be considered. If the coverage is not good, it may not be able to locate indoors in some places because the signal cannot be received.This paper is different from the previous indoor positioning methods. It is an indoor positioning paper that does not rely on Bluetooth and WIFI devices and relies entirely on computer vision. The model of this paper is based on the existing and mature object recognition and extends its function with a model of depth estimation and angle estimation to locate the user's location information in the room, which can facilitate the user's operation. For example, users can directly pick up a smartphone and scan any object in the museum with a camera to know the user's location. They can also make extended applications such as path planning for museum tours.

第一章 緒論 1
第二章 文獻探討 3
第三章 研究方法 15
第四章 訓練資料實驗過程與結果 23
第五章 結論與未來展望 38

[1] M. D'Aloia et al., "IoT Indoor Localization with AI Technique," 2020 IEEE International Workshop on Metrology for Industry 4.0 & IoT, 2020, pp. 654-658, doi: 10.1109/MetroInd4.0IoT48571.2020.9138275.
[2] Neocognitron, et al. “Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position” Biological Cybernetics volume 36, pages193–202 (1980)
[3] Y. Lecun, L. Bottou, Y. Bengio and P. Haffner. (1998). Gradient-based learning applied to document recognition. in Proceedings of the IEEE, vol. 86, no. 11, pp. 2278-2324, Nov. 1998, doi: 10.1109/5.726791.
[4] Alex Krizhevsky, et al.” One weird trick for parallelizing convolutional neural networks” arXiv:1404.5997[cs.CV] (2014)
[5] Karen Simonyan, Andrew Zisserman ” Very Deep Convolutional Networks for Large-Scale Image Recognition” arXiv:1409.1556 [cs.CV] (2014)
[6] Christian Szegedy, Wei Liu, Yangqing Jia, Pierre Sermanet, Scott Reed, Dragomir Anguelov, Dumitru Erhan, Vincent Vanhoucke, Andrew Rabinovich ” Going Deeper with Convolutions” arXiv:1409.4842 [cs.CV] (2014)
[7] Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun ” Deep Residual Learning for Image Recognition” arXiv:1512.03385 [cs.CV] (2015)
[8] Ross Girshick, Jeff Donahue, Trevor Darrell, Jitendra Malik ” Rich feature hierarchies for accurate object detection and semantic segmentation” arXiv:1311.2524 [cs.CV] (2014)
[9] Joseph Redmon, Santosh Divvala, Ross Girshick, Ali Farhadi ” You Only Look Once: Unified, Real-Time Object Detection” arXiv:1506.02640 [cs.CV] (2016)
[10] Baichuan Huang, Jun Zhao, Jingbin Liu ” A Survey of Simultaneous Localization and Mapping with an Envision in 6G Wireless Networks” arXiv:1909.05214 [cs.RO] (2020)
[11] David Eigen, Christian Puhrsch, Rob Fergus ” Depth Map Prediction from a Single Image using a Multi-Scale Deep Network” arXiv:1406.2283 [cs.CV] (2014)
[12] Iro Laina, et al.” Deeper Depth Prediction with Fully Convolutional Residual Networks” arXiv:1606.00373 [cs.CV] (2016)
[13] Divyanshu Mishra” Transposed Convolution Demystified”
:https://towardsdatascience.com/transposed-convolution-demystified-84ca81b4baba
[14] J. Deng, W. Dong, R. Socher, L. -J. Li, Kai Li and Li Fei-Fei, "ImageNet: A large-scale hierarchical image database," 2009 IEEE Conference on Computer Vision and Pattern Recognition, 2009, pp. 248-255, doi: 10.1109/CVPR.2009.5206848.