Download as pdf or txt
Download as pdf or txt
You are on page 1of 4

Proceedings of International Conference on Circuits, Communication, Control and Computing (I4C 2014)

An Image Processing Technique for Color Detection


and Distinguish Patterns with Similar Color: An aid
for Color Blind People
Bhagya R Navada1, Santhosh K V2, Prajwal S, Harikishan B Shetty
Dept. of Instrumentation and Control Engineering
Manipal Institute of Technology
Manipal, India
1
kgbagya@gmail.com, 2kv.santhu@gmail.com

Abstract— In this work an attempt is made for designing a ear detection by edge detection and template matching
prototype in aid for color blind people in detecting color and approach is reported in [4] for biometric application. In [5],
edges of a given image which are of similar colors. Color blind measurement of distance between two objects on an oblique
people are restricted and even excluded from some professional plane is reported using image based frame work. An
jobs because of their deficiency to distinguish between colors.
application of gradient descent method for image segmentation
Motivation behind this work is to try and reduce such deficiency
using technology. Image processing technique is used in proposed of optic nerves of a bird’s eye is discussed in [6]. Here Signed
work to develop a setup for identifying the colors in an image and Pressure Force function is used for initialization of local
indicating the edges in case of different images with similar contour. An algorithm is used in [7] for edge detection of a
colors. Image processing involve two parts, first is color detection color image by using threshold technique. In [8], a method is
and the second is edge detection which is carried out on a proposed for recursive detection of edges in color image based
LabVIEW platform. Images are obtained using a wireless on Green’s function approach in the color vector space field. A
camera in the field of view of a color blind person who is unable technique is reported in [9] for color detection using
to identify or judge the colors. It is then processed using hierarchical neural networks in RGB space. A method for color
LabVIEW, color and edge of the image is displayed on LabVIEW
image edge detection using fast vectorial total variation
front panel.
minimization model is reported in [10]. In [11], a color
Index Terms—Image processing, color blindness, LabVIEW. recognition system is developed by finger interaction method
Edge detection for object color detection. A traffic light recognition system is
developed in [12] using image processing techniques for color
I. INTRODUCTION blind vehicle driver.
Out of the varied spectrum of vibrations, only a tiny part of From the study of reported work it is clear that, very less
it is registered by the optic nerves. The perception of color is a work is done in the area of designing aid for color blind people.
sensory experience. The color of an object is seen by the eye Our aim is to develop an aid for color blind people for
when white light is shone upon the object's surface. Color recognizing color and the edges of images. The proposed setup
blindness is the inability to differentiate among certain colors. involves wireless camera through which the image will be
Approximately 8% of men and 1% of women have a form of acquired, TV tuner card for interfacing the camera with
color vision deficiency, which can be genetic or acquired computer. The acquired image will be processed using the
through illness, accident or poisoning. It is often a life-long LabVIEW IMAQ vision and vision assistant. The acquired
condition. Color plays an important role in all kinds of image and the processed image with colors will be displayed
industries. Along with color, an object can be characterized by on the front panel of VI.
its size and shape. Edges are significant local changes of LabVIEW gives opportunity to develop code which could
intensity in an image. Edges typically occur on the boundary differentiate between colors.
between two different regions in an image. By determining its IMAQ Vision for LabVIEW is a library of LabVIEW VIs
edges, an object’s shape and size can be identified. There are that can be used to develop machine vision and scientific
many applications of edge detection like surveillance, text imaging applications.
detection and shape detection etc. Of all the different image formats that the Vision Toolkit
Many researchers have reported work on image processing supports by default, JPEG (Joint Photographic Experts Group)
techniques like, in [2], a method is reported for contrast are probably the most common. The JPEG format is optimized
improvement of typhoon cloud images by using a normalized for photographs and similar continuous tone images that
arc tangent transformation and differential evolution algorithm. contain a large number of colors, and can achieve astonishing
An image segmentation technique is discussed in [3] to check compression ratios, even while maintaining a high image
the ripe tomatoes using K-means clustering algorithm. Human quality. The JPEG compression technique analyzes images,

MSRIT, BANGALORE, India, 21-22 NOVEMBER 2014 978-1-4799-6546-5/14/$31.00 © 2014 IEEE 333
Proceedings of International Conference on Circuits, Communication, Control and Computing (I4C 2014)

removes data that is difficult for the human eye to distinguish, III. PROBLEM DESCRIPTION
and stores the resulting data as a 24-bit color image [1]. Aim of this work is to design an aid for people suffering
Paper is organized as follows: After introduction in section from color blindness. In this work image is captured from a
I, section II describes the experimental model. Problem wireless camera. Output of the camera is digitized by TV tuner
description is given in section III, Methodology in section IV card which is connected to the computer through USB.
result analysis is discussed in section V and in section VI Acquired image is processed using LabVIEW IMAQ vision
conclusion and future work is discussed. tools. Processing involves two parts.
• Color detection
II. EXPERIMENTAL SETUP • Edge detection.
Fig. 1 shows the experimental setup used for proposed IV. METHODOLOGY
work. It consists of wireless camera, TV tuner card, computer
Block diagram of the proposed work is shown in Fig. 2
with LabVIEW software.
A. Wireless A/V Camera: Wireless camera with transmitter
Wireless cameras transmit a video and audio signal to a
wireless receiver through a radio band. Wireless cameras are
proving very popular among modern security consumers due to
their low installation costs and flexible mounting options; Receiver
wireless cameras can be mounted/installed in locations
previously unavailable to standard wired cameras.
Digitizer (TV tuner card)

Computer with LabVIEW


Fig. 2 Flow diagram
Using wireless camera, live image is acquired. The input
image is digitized using TV tuner card & digitized image is
processed using LabVIEW to find color and edges of image.
The programs were developed using the 2011 version of
LabVIEW. Given below are some of the tools that were
helpful in developing the program:
• IMAQ Create: Creates a temporary memory location
for an image.
• IMAQdx Open Camera VI: Opens a camera, queries
the camera for its capabilities, loads a camera
configuration file, and creates a unique reference to the
camera.
• IMAQdx Snap VI: Configures starts, acquires, and
unconfigures a snap acquisition.
• IMAQdx Grab VI: Acquires the most current frame
into Image Out. Call this VI only after calling
Fig. 1 Experimental setup
• IMAQdx Configure Grab VI: If image type does not
B. Digitizer (TV Tuner card) match video format of camera, this VI changes image
type to a suitable format.
A TV tuner is a kind of television tuner that allows
television signals to be received by a computer. Most TV • IMAQdx Close Camera VI: Stops an acquisition in
tuners also function as video capture cards, allowing them to progress, releases resources associated with an
record television programs onto a hard disk much like the acquisition, and closes the specified Camera Session.
digital video recorder (DVR) does. • Index Array: Returns element or sub array of n-
The interfaces for TV tuner cards are most commonly dimension array at index.
either PCI bus expansion card or the newer PCI Express bus • IMAQ Extract Single Color Plane VI: Extracts a single
for many modern cards, but PCMCIA, Express Card, or USB plane from a color image.
devices also exist. In addition, some video cards double as TV • IMAQ Edge Detection VI: Extracts contours (detects
tuners, notably the ATI All-In-Wonder series. The card edges) in gray-level values. Any image connected to
contains a tuner and an analog-to-digital converter along with the input Image Dist. must be the same image type
demodulation and interface logic. connected to Image Src. The image type connected to

MSRIT, BANGALORE, India, 21-22 NOVEMBER 2014 334


Proceedings of International Conference on Circuits, Communication, Control and Computing (I4C 2014)

the input Image Mask must be an 8-bit image. The • Enhancement: apply a filter to enhance the quality of
connected source image must have been
b created with a the edges in the image (sharpening).
(
border capable of supporting the size
s of processing • Detection: determine which edge pixels should be
matrix. discarded as noise and which
w should be retained.
• IMAQ Color Match VI: Finds the maatch between color • Localization: determinee the exact location of an edge.
content of multiple regions in an image and those Edge thinning and linkking are usually required in this
defined by the input color Spectrum. Color Spectrum is step.
output from the IMAQ Color Learn VI.V
• Region of Interest - A group of VIs that
t manage ROIs. V. RESULTS AND
A ANALYSIS
Use these Vis to programmatically define ROIs and Our aim is to find colorr and edge of a given image
convert ROIs to and from image massks [13]. according to the region of innterest. By running this VI, a
picture input is read from cameera. Select the region of interest
Edge detection is a fundamental tool ussed in most image in picture whose color is to be found.
f The VI displays the color
processing applications to obtain informatioon from the video of selected points in the front panel
p as shown in fig. 3. IMAQ
frames in which feature extraction and object segmentation is color match is used for detectinng the color after training it to n
done In the ideal case, result of applying an edge
e detector to an colors.
image may lead to a set of connected curvees that indicate the
boundaries of objects, boundaries of surfacee markings as well
as curves that correspond to discontinuuities in surface
orientation. Thus, applying an edge detectioon algorithm to an
image may significantly reduce the amouunt of data to be
processed and may therefore filter out informmation that may be
regarded as less relevant, while preserviing the important
structural properties of an image. If edge detection step is
successful, subsequent task of interpretingg the information
contents in original image may thereforee be substantially
simplified. However, it is not always possibble to obtain such
ideal edges from real life images of modderate complexity.
There are an extremely large number of o edge detection
operators available, each designed to be seensitive to certain
types of edges. Variables involved in seleection of an edge
detection operator include:
Extracts contours (detects edges) in grayy-level values. The
image type connected to input Image Maskk must be an 8-bit
image. The connected source image must have h been created Fig. 3 Color detecction of a given image
with a border capable of supporting the sizee of the processing From an input image, 3 coolors can be identified from the
matrix. database of 7 colors present coomprising of violet, indigo, blue,
Noise environment: Edge detection is difficult in noisy green, yellow, orange and red.
images, since both the noise and the eddges contain high Acquired image is which is shown in fig. 4 is processed and
frequency content. Attempts to reduce thhe noise result in the edges of filtered image is deetected which is shown in fig. 5
blurred and distorted edges. Morphology VI does perform
basic morphological operations, such as dilation and erosion,
on grayscale and binary images. Other VI im mprove the quality
of binary images by filling holes in paarticles, removing
particles that touch the border of an imagee, removing noisy
particles, and removing unwanted particles based
b on different
characteristics of the particle.
Edge structure: Not all edges involve a step change in
intensity. Effects such as refraction or poor focus
f can result in
objects with boundaries defined by a grradual change in
intensity. The operator needs to be chosen too be responsive to
such a gradual change in those cases. Sobel algorithm
techniques actually characterize the nature of the transition for
each edge in order to distinguish.
Steps of edge detection
• Smoothing: suppress as much nooise as possible,
without destroying the true edges Fig. 4 Origginal Image

MSRIT, BANGALORE, India, 21-22 NOVEMBER 2014 335


Proceedings of International Conference on Circuits, Communication, Control and Computing (I4C 2014)

work can also be extended by converting the recognized text


into speech and sending that to the blind people.

REFERENCES
[1] Christopher. G. Relf, ‘Image acquisition and processing with Lab
VIEW’, CRC Press LLC, 2004
[2] Chang-Jiang Zhang, Bo Yang, “A Novel Nonlinear Algorithm for
Typhoon Cloud Image Enhancement”, International Journal of
Automation and Computing, Vol. 8, No. 2, May 2011, 161-169.
[3] Monika Bhatnagar, Prashant Kumar Singh, “An Efficient Method of
Image Segmentation for Harvest Time Identification”, International
Journal of Computer Applications, vol 87, No.7, February 2014.
[4] K. V. Joshi, N. C. Chauhan, “Edge Detection and Template Matching
Approaches for Human Ear Detection”, International Journal of
Computer Applications Special Issue on Intelligent Systems and Data
Processing, 2011, 50-55,
[5] Chen-Chien Hsu, Wei-Yen Wang, and Yin-Yu Lu, Ming-Chi Lu,
“Image-Based System for Measuring Objects on an Oblique Surface”,
Instrumentation and Measurement Technology Conference (I2MTC),
IEEE, 2010, 313 - 317
Fig. 5 Edge detection of a given image [6] Pushpendra Kumar, Rekha Pandit, Vineet Richhariya, “Retinal Image
Segmentation by using Gradient Descent”, International Journal of
Computer Applications, Volume 86 – No 10, January 2014, 1-7.
The system was subjected to test with images having
[7] Soumya Dutta, Bidyut B. Chaudhuri, “A Color Edge Detection
different color combinations. 64 such images was tested with Algorithm in RGB Color Space”, International Conference on Advances
the proposed algorithm, result produced was matching actual in Recent Technologies in Communication and Computing, 2009, 337 -
color in all the 64 cases. Similarly to check its functionality 340.
images with text having similar background color (e.g. Light [8] Zahra Zareizadeh, Reza P.R. Hasanzadeh∗, Gholamreza Baghersalimi,
blue background and text in dark blue) was also used to “A recursive color image edge detection method using Green’s function
approach”, Optik - International Journal for Light and Electron Optics,
identification. Make it convenient for the color blind person to vol. 124, Issue 21, 2013, 4847–4854
detect and read text easily. [9] Halis Altun, Recai Sinekli, Ugur Tekbas, “An Efficient Color Detection
in RGB Space Using Hierarchical Neural Network Structure”,
VI. CONCLUSION AND FUTURE SCOPE OF WORK International Symposium on Innovations in Intelligent Systems and
Objective of the work is satisfied by the proposed method Applications (INISTA), 2011, 154 – 158.
of finding color and edges of a given image. LabVIEW IMAQ [10] Jun Zeng∗, Dehua Li, “Color image edge detection method using VTV
denoising and color difference” Optik - International Journal for Light
vision and vision assistant are used as a development tool for and Electron Optics, vol. 123, Issue 22, 2012, 2072–2075.
finding the color and edge of a color image. The whole setup of
[11] Alfa Sheffildi Manaf, Riri Fitri Sari, “Color Recognition System with
the work is found to be cheap, feasible, flexible & effective. Augmented Concept and Finger Interaction”, Ninth International
When the letters are written in some colors with some other Conference on ICT and Knowledge Engineering, 2011, 118 - 123
background color, it will be difficult for the color blind to [12] Youn K. Kim, Kyoung W. Kim, Xiaoli Yang, “Real Time Traffic Light
recognize the letters. Future work of the proposed method is to Recognition System for Color Vision Deficiencies”, Proceedings of the
2007 IEEE International Conference on Mechatronics and Automation,
detect the text by interfacing LabVIEW with the GSM module
August 5 - 8, 2007, Harbin, China.
which is used to send processed information to user thus
[13] Thomas Klinger, Image Processing with LabVIEW and IMAQ Vision,
helping the color blind people in recognizing the text. This Prentice Hall Professional, 2003.

MSRIT, BANGALORE, India, 21-22 NOVEMBER 2014 336

You might also like