CSEIT183116 | Received : 01 Jan 2018 | Accepted : 12 Jan 2018 | January-February-2018 [(3) 1 : 97-101 ]
International Journal of Scientific Research in Computer Science, Engineering and Information Technology
© 2018 IJSRCSEIT | Volume 3 | Issue 1 | ISSN : 2456-3307
97
Image Processing Based Bacterial Colony Counter
Bhavika Jagga
*
, Dr. Dilbag Singh
Department of Computer Science and Applications, Chaudhary Devi Lal University, Sirsa, Haryana, India
ABSTRACT
Enumeration of Bacterial Colonies is required in many fields such as in clinical diagnosis, biomedical research for
prevention of harmful diseases and pharmaceutical industry to avoid contamination of products. Existing Bacterial
Colony counter systems count Bacterial Colony manually which is a time consuming, less efficient and tedious
process. Hence, automation for counting of bacterial colony was required. The proposed method count these
colonies automatically using image processing techniques. This method will provide a greater degree of accuracy in
counting of bacterial colonies. Proposed technique takes an image of bacterial colony and converts it into grayscale.
Otsu thresholding is applied for segmentation of the image further its conversion into binary image. After that,
morphological operations are applied to clean up the image by removing noise and unnecessary pixels. Distance and
watershed transformations are applied on the binary image to create partitions among overlapped and joint bacteria.
Region properties and labeling information of segmented image is used for counting of bacterial colony.
Keywords: Bacterial Colony, Thresholding, Morphology, Distance Transform and Watershed Segmentation.
I. INTRODUCTION
Bacterial Colony is defined as a cluster of bacteria
derived from one common bacterium. Microbiologists
require accurate measure of the bacterial colonies for
many biological procedures [1]. Enumeration of
Bacterial Colonies is important for obtaining precise
assessment of pathogens. Manual counting of bacterial
colonies is a tedious and time consuming process.
Automation of the process of counting bacterial
colonies will save time and labor required for counting
of colonies [2]. Image segmentation techniques are
used to automate this counting process. Image
segmentation changes the image into a form suitable
for image analysis. In image segmentation, a digital
image is divided into multiple segments. Image
segmentation assigns label to each pixel of an image on
the basis of visual characteristics. Image segmentation
is very used in many areas such as object detection,
image retrieval, object-based counting, tissue
identification, cell counting and object tracking [3].
Watershed segmentation is widely used for counting
microorganisms. It is classified as region-based
segmentation approach. It is used to separate two
touching objects or overlapping objects. It considers
image as a topographic surface in which the gray level
of each pixel is considered to be the height on the
surface and the two overlapping or touching objects as
catchment basins. The motive is to find out watershed
ridge lines separating the two catchment basins for
separating the two overlapping objects [4] [5].
II. LITERATURE REVIEW
Sethi and Yadav (2012) used the multi- threshold
segmentation procedures to count bacterial colonies for
separating and detecting the colonies present. Final
processed image is used for counting of separated
colonies using a conventional single-threshold
segmentation procedure. Results depicted the low and
medium density bacterial colonies. It have been
observed that the proposed technique does not hold
good for low contrast images and high density medium
of bacterial colonies. In case of low contrast images of
colonies gets distorted after thresholding, leading to
appearance of high curvature points along the
boundary. These high curvature points get accumulated
in count result [1].