基於智慧型手機自動菌落計數系統

Mobile Phone Based Bacteria Colony Counting

張証凱(Cheng-Kai Chang)

指導教授 ： 黃乾綱

國立臺灣大學/工學院/工程科學及海洋工程學研究所/碩士(2013年)

https://doi.org/10.6342/NTU.2013.00438

摘要

菌落計數是微生物實驗的基礎及必要的工作，然而在實驗的過程中，最令實驗者困擾的地方在於培養實驗結束後，必需估算為數眾多培養皿內的菌落數，這一個非常重要但是相當耗費人力時間的步驟。為了解決上述問題，市面上已經有商業化的菌落自動計數機器，但是這些儀器價格上十分昂貴，使用上無法任意搬移到特定地區。隨著智慧型手機的普及率漸漸提升，本研究開發一款手機APP來解決上述問題，APP透過手機的攝影鏡頭擷取培養皿內的菌落數影像，再利用影像分析技術與機器學習方法技術，統計出菌落的數量。本研究特別針對智慧型手機APP設計演算法，避免在處理計算菌落的時候發生記憶體不足的情況。相較於其他相關研究所提出的方法，本研究方法不僅快速、準確率達97%且便利，更可以解決更多種不同顏色的菌落與更多不同顏色的背景的培養皿圖像。同時本研究也提供了網路服務，使用者可以利用APP上傳培養皿圖片到此服務進行大量且快速的自動菌落計數。

關鍵字

菌落計數

並列摘要

Microbiological experiment is the basis of biotechnology. Colony counting is a very important step after experiments. However colony counting is a time-consuming and inefficient process in the microbiological experiments. In this thesis, with the population of smartphone, we proposed an approach and developed a mobile application, which utilize the mobility of smartphone to help user count the number of colonies. After taking photos of petri dish, machine learning and image processing are used to count the number of colonies. Compare with previous work, our approach overcome the difficulties of processing various type of colonies and culture medium. We also propose a new approach which prevent the out of memory situation while the smartphone is processing the colony images. Each pixel in the image of colony is viewed as a data. The data is going to cluster with K-means algorithm. Thereafter three features are extracted including area, perimeter and shape factor. These features are used to identify colonies. We also provide an on-line service to which user can use the APP to upload images of petri dish for mass and rapid automatic colony counting. Our approach is not only fast and convenient but also capable to deal with various type of colony with various backgrounds with the precision rate 97%.

並列關鍵字

Colony count

參考文獻

2. Men, H., et al. Application of support vector machine to heterotrophic bacteria colony recognition. 2008. IEEE.

3. Men, H., et al. Counting method of heterotrophic bacteria based on image processing. in Cybernetics and Intelligent Systems, 2008 IEEE Conference on. 2008. IEEE.

4. Wei-zheng, S., et al. Experimental study for automatic colony counting system based on image processing. 2010. IEEE.

5. Cortes, C. and V. Vapnik, Support-vector networks. Machine learning, 1995. 20(3): p. 273-297.

7. Vincent, L. and P. Soille, Watersheds in digital spaces: an efficient algorithm based on immersion simulations. IEEE transactions on pattern analysis and machine intelligence, 1991. 13(6): p. 583-598.