自20世紀以來，電腦與網路的高速進步促使以其為中心衍生的技術大幅發展，其中ai的相關研究如機器學習優化了人類在資料處理上的相關應用。本研究欲透過結合大數據與機器學習的模式，在智慧建築的領域提供一條可行性的方向討論，透過影像辨識的技術拓展現今數位雙生中資料蒐集的可能性。
本研究以(1)定義平面空間領域人流「場域辨識感知」、(2)影像資料整合進行非監督式學習的「機器學習分群」、以及(3)實證分析的「資料視覺化」三個範疇來詮釋這套人流大數據資料分析系統的架構以及實證研究的方向。透過opencv影像處裡、yolov5影像辨識、以及kmeans機器學習資料分群等代碼建構整個系統所需要的環境及骨幹。
本研究希望藉由電腦資料處裡的角度分析在建築中人流行為的模式是否與人眼所見的既定模式有差異性可以做討論。另外、再未來能導入此研究系統協助一些智慧建築維運等的智慧化控制決策，藉由當下人流狀態分析去完成一些物理環境控制等智慧建築管理。

Since the 20th century, the rapid advancement of computers and networks has significantly propelled the development of technologies centered around them, notably the field of artificial intelligence, which has optimized human data processing applications. This study aims to explore a feasible direction within the domain of smart buildings by integrating big data and machine learning patterns, expanding the possibilities of data collection within the modern concept of digital twins through the technology of image recognition.

The research delineates the architecture and the empirical study direction of a pedestrian traffic big data analysis system within three domains: (1) "Spatial Domain Recognition and Perception," which defines the pedestrian traffic in a planar space; (2) "Unsupervised Learning Machine Learning Clustering," which involves integrating image data for unsupervised machine learning; and (3) "Data Visualization," which interprets the data through demonstrative analysis. The system's environment and infrastructure are constructed using codes from OpenCV image processing, YOLOv5 image recognition, and K-Means machine learning clustering.

This study aspires to analyze pedestrian behavior patterns within architecture from the perspective of computer data processing to discuss potential deviations from established patterns observed by the human eye. Additionally, the study envisions future incorporation of this research system to assist smart architectural operations and intelligent control decisions. By analyzing the current state of pedestrian flow, the system could facilitate various physical environment controls and the management of smart buildings

中文文獻
1.廖士豪（2020）。整合AI電腦視覺與BIM電子圍籬發展智慧維運平台〔碩士論文，逢甲大學〕。
2.楊佳恩(2018)。建築資訊模型雲端視覺化系統應用於智慧維運之研究。逢甲大學建築所，台中市。
3.張宗彥 ( 2023 )。OPENCV與影像辨識。台北市:深智數位
4.簡琳儒. (2020, December 29). 影像辨識涵構察覺應用於群眾行為之空間關聯性探討-以台中文華路為例，成功大學建築所，台南市
5.張宗彥 ( 2023 )。OPENCV與影像辨識。台北市:深智數位
6.盧建成譯 ( 2020 )。非監督式學習-使用Python ( 原作者 : Ankur A. Patel )。美國 : 歐萊禮出版社。(原著書名 : Hands-on Unsupervised Learning Using Python)
外文文獻
1.Hinton, G. E., Osindero, S., & Teh, Y. W. (2006). A fast learning algorithm for deep belief nets. Neural computation, 18(7), 1527-1554.
2.El Naqa, I., & Murphy, M. J. (2015). What is machine learning? (pp. 3-11). Springer International Publishing.
3.Rosenblatt, F. (1958). The perceptron: A probabilistic model for information storage and organization in the brain. Psychological Review, 65(6), 386-408.
4.Breiman, L., Friedman, J. H., Olshen, R. A., & Stone, C. J. (1984). Classification and regression trees. Wadsworth & Brooks/Cole Advanced Books & Software.
5.Hinton, G. E., Osindero, S., & Teh, Y. W. (2006). "A fast learning algorithm for deep belief nets." Neural Computation, 18(7), 1527-1554.
6.Jain, A.K. (2010). "Data clustering: 50 years beyond K-means." Pattern Recognition Letters, 31(8), 651-666.
7.Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A., & Bengio, Y. (2014). "Generative Adversarial Nets." In Advances in Neural Information Processing Systems.
8.Liu, B., Yan, J., & Zhou, D. (2017). "A survey of unsupervised learning models and algorithms." Expert Systems with Applications, 55, 412-424.
9.Yan, J., Meng, Y., Lu, L., & Li, X. (2020). "Building Structure Analysis Based on Unsupervised Learning." Journal of Civil Engineering and Management, 26(8), 763-775
10.Hotelling, H. (1933). Analysis of a complex of statistical variables into principal components. Journal of educational psychology, 24(6), 417.
11.Paea, S., & Baird, R. (2018). Information Architecture (IA): Using multidimensional scaling (MDS) and K-Means clustering algorithm for analysis of card sorting data. Journal of Usability Studies, 13(3), 138-157.
12.Hinton, G. E., Osindero, S., & Teh, Y. W. (2006). A fast learning algorithm for deep belief nets. Neural computation, 18(7), 1527-1554.
13.Chen, T. W., & Chien, S. Y. (2009). Bandwidth adaptive hardware architecture of K-means clustering for video analysis. IEEE transactions on very large scale integration (VLSI) systems, 18(6), 957-966.
14.Chouinard, J. C. (2023, December 25). What Is KMeans Clustering Algorithm (with Python Example) – Scikit-Learn. JCCHOUINARD.