Improved Performance of Fuzzy Logic Algorithm for Lane Detection Images

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 04 Issue: 07 | July -2017 www.irjet.net p-ISSN: 2395-0072
© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 3151
Improved Performance Of Fuzzy Logic Algorithm For Lane Detection
Images
Tamanna1 , Arushi Bhardwaj2
1M.Tech Scholar, Department of Electronics & Communication Engineering,
Sri Sai College Of Engineering & Technology, Badhani, Pathankot, Punjab, India
2Assistant Professor, Department of Electronics & Communication Engineering,
Sri Sai College Of Engineering & Technology, Badhani, Pathankot, Punjab, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract- This paper represents that lane coloration has become popular in real time vehicularadhocsystem.Lane
detection is normallyhelpful to localize path limits. Determine undesired lanevariationsandtoenableapproximation
of the upcoming geometry of the road. There are different types of methodsthat are used for detecting lines i.e.Hough
transform, clustering and curve fitting. The paper shows that Hough transform, clustering and curve fitting work
efficiently, but problem is that may fail or not give efficient results when there are curved lane road images. The
objective of this paper is to improve lane coloration algorithm by modifying the Hough transform i.e. fuzzy logic with
different performance metrics to improve the accuracy. Extensive experiments have shown that proposed technique
outperforms over the available techniques.
Key words: Lane detection, Hough transform and Fuzzy logic.
1.Introduction
Passenger's safety is probably one of the most formulated axes concerning exploration in automobile. The vastmajorityofthe
vehicle road crashes takes place because of the driver overlookingofthevehiclepathsoprotectionistheprimarypurposeofall
of the lane detection methods. The majority of travelling deaths and injuries happen on the country’s highways. According to
the fact, improper driving response, high speed as well as U-turn are the main causes behind majority of these incidents.
Studies of these accident cases depict that 40% and more mishaps could have been eliminated if perhaps the vehicle hadbeen
designed with an alert system. The next generation of driver-assistant system are being developed by consumer analysis
organizations, automobile manufacturers and suppliers, as well as other research institutions that will make it possible for
vehicles to have more secure tendencies as well as to decrease road injuries and deaths. A computer perspectiveisinvolvedas
one of the primary technologies which become a powerful tool for detection of lanes [18].Lanedetectionisnormallyhelpful to
localize road boundaries, determine undesired lane variations, and to enable approximation of the upcoming geometry of the
road. At Present, two well defined techniques are there for performing lane recognition by making use of video i.e. feature
based method and model based method [1, 9] Lane detection enables you to obtain the position as well as direction of the
vehicle in addition lane information, as well as an area which includes highways is important to alert a driver associated with
lane departure. The lane information is usually used for tracking down othermotorvehiclesaswell ashurdleswithintheroute
of the vehicle and which could be placed on additional growth of the barrier avoiding system [7].
1.1 Hough Space:
There are various ways of representationofstraightlinesin2Dcoordinatesystems.InCartesiancoordinatesystem
(CCS) straight line can be represented with the help of following equation:
y=m*x+c (1)
Where m is slope, b is the intercept on y axis. This can also be represented with the help of figure 1.9 [4]

Figure 1. Line in Cartesian coordinate system
This is not a proper way of representation of straight lines, as vertical lines cannot be represented using this
equation. This is because slope of line becomes infinite. So we use polar coordinate system for representation of
straight lines for parameters (r,θ). Straight line using polar coordinate system can be represented with the help of
following equation [4]:
y = ) *x+( (2)
Where r = xcosθ+ysinθ (3)
Here r is the length of perpendicular to this line, starting from origin and θ is the orientation angle ofrwithrespect
to x axis. Figure 1.10 shows a line in polar coordinate system [4].
Figure 2.Line in polar coordinate system
1.2 Hough Transform:
Analysis of detecting lines, curves and ellipses is globally done by Hough transform techniques. It is generally
applied after performing edge Detection. According to Hough Transform “Every single pixel in image space
corresponds to a line inside a parameter space” also called hough space [18]. The Hough transform(HT)isauseful
resource for detecting straight lines inside pictures, even presence associated with noise andmissinginformation,
becoming a trendy choice for this task [2]. The Hough transform, HT, was presented as a technique of sensing
complicated factors in binarypictureinformation[1].Itdefinesthatbydecidingparticularpricesofvariableswhich
characterize these patterns. Spatially lengthy habits are altered so they make spatially lightweight functions in a
place of probable parameter values. The HT switches an recognition issue in picture place right into a quicker
resolved regional top recognition issue inaparameterspace.Themainelementsomeideasofthetechniquemaybe
shown by contemplating distinguishing models of collinear factors in a image. A couple of picture factors (x, y)

which lay on a direct point may be described by way of a connection, f, in a way that f((vh, e), (x, y)) = b – rhx = 0,
(1).
Hough Transform Algorithm:
Require: {Binary Image}
Require: δ {Discretization step for the parameter space}
1: Votes ← 0 {Initialization of the voting matrix}
2: for each feature pixel I (x, y) do
3: for 0°≤ Ѳ ˂ 180°, using a δ Discretization step do
4: ρ ← x cos(Ѳ) + y sin(Ѳ)
5: Votes (ρ, Ѳ) ← Votes (ρ, Ѳ) + 1
6: end for
7: end for
1.3 Fuzzy Logic:
Fuzzy logic idea is usually than the man being's sensation in addition to inference process. As opposed to
established management approach, that really is a point-to-point management, hairy judgement management is a
range-to-point and also range-to-range control. The particular creation of any hairy controlled comes from
fuzzifications associated with either inputs in addition to components making use of the connected member
functions. The crisp and clean feedback might be changed to unique members of the particular connectedmember
features predicated upon it's value. Using this point of view, the particular creation of any hairy judgement
controlled can be launched upon it's memberships of various member features, that is thought of as several
different inputs.To implement fuzzy logic technique to a real application requires the following three steps:
1. Fuzzifications – convert common information or perhaps highly detailed information directly into hairy
information or perhaps Member Functions (MFs).
2. Fuzzy Inference Process – put together account features while using the manage procedures so that you can get
the particular unclear output.
3. Defuzzification - apply several methods to estimate every associated outcome plus insert them in tothetable:yourresearch
table. Grab your outcome in the research table on such basis as today's input during an application.
Fig 6: Fuzzy Controller [9]
Preprocessing Fuzzification
Rule Base
Inference
Engine
Defuzzification PostProcessing

2.Related work
Qing Lin et al. [1] proved that the lane colorizationcanpresentsignificantinformationforprotectiondriving.Areal
time vision-based lane colorization process has been offered to locate the location and form of lane in every video
structure. In lane colorization process, lane assumption has been generate and established based on an efficient
grouping oflane-markboundary-linkfeatures.TatsuyaKasaietal.[2]verifiedthelanemarkerdetectionmethodfor
platooning. For decreased the air resistance, it has been attractivetowardcutdownthespaceamongtwovehicle.If
the vehicular space has very small, conservative process, which identifieslanesymbolsinimageryimprisonfroma
frontage camera, has been in efficientforthereasonthatlaneindictorshasoccludedviaavehicleinfront.Chunzhao
Guo et al. [3] Proved that the concurrent lane recognition and localization has been a single key issue for several
intelligent transport systems. It has defined lane recognition and tracking process to work in difficult situation
where lane border scan be low-distinction and changeful with sound due to a number of factors such as kind,
illumination and climate environment, etc. Jia He et al. [4] Described that the visualization based Lane Departure
Warning System has been an efficient method to avoid Single auto mobile high way Departure mishap. In
performance, a range of composite sound make it extremelyhard to identifylanequicklyandcorrectly,sotocreate
a variety of picture processing technique which can provide out come quickly and correctly in the non-ideal
environment has been the main exertion. Yu-Chi Leng et al. [5] proposed that the lane-departure recognition
method with no fundamental and extrinsic camera parameter calibration.Itprovidesdrivingprotectionwithtrack
recognition and track departure warning has been paying attention on metro polite an highway through complex
track symbols as an alternative of straight forward road scene. Yong Chen et al. [6] described the efficient lane
borders projective model and enhanced recognition process in the image capture by a vehicle-mount monocular
camera in composite environment, for point rounded arc path initially, a lane borders projective model has been
assumed. This lane model not only describes straight-line lane borders, although besides express the tangible
pointed rounded arc path borders extremely clearly. Hendrik Deusch et al. [11] proposed that the robust lane
colorizationhasbeentheprerequisiteusedforsophisticateddriversupportsystemsimilartolanedepartureadvice
and overtake subordinate . [18]Jung, Soonhong, Junsic Youn, and Sanghoon Sull, proposed a proficient solution to
reliably detectingpathlanesdependingonspatiotemporalimages.Inalinedupspatio temporalimagegeneratedby
accumulating the actual p on the scan line along the time axis plus aiming continuous scan lines, the actual
trajectory of the side of the road items looks sleek plus styles some sort of direct line. A aligned spatiotemporal
photo can be binarized, and 2 principal simultaneous straight lines resultingfromtheactualtemporaryuniformity
involving side of the road size on the given scan line usually are noticed having a Hough transform, cutting down
stance errors. The everywhere you look side of the road items usually are in that case noticed nearby the
intersections of the actual direct outlines along with the latest scan line. Our spatiotemporal domain approach can
be tougher lost or maybe occluded lanes as compared with existing frame-based approaches.

3.PROPOSED METHODOLOGY
Fig 1: Proposed Methodology
4.Experimental results
For experimentationandimplementationtheproposedtechniqueisevaluatedusingMATLABtoolu2013a.Herewe
compare the lane colorization algorithm i.e. Hough transform with additive Hough transform for removing noise
from the images on the basis of various image quality evaluation parameters like recall, f-measure, p_recallandbit
error rate. The existing methodology give good results which locate your ln corners with virtually no previous
expertise on the highway geometry, in addition to do it within cases for you might be a many chaos inside the
highway photograph Thus it becomes a major issue when noise is present in the input image. The proposed
approach gives efficient results in improving the existing Lane colorization algorithm. The particular tabular in
addition to graphic contrast is accomplished in between established in addition to consist of method based on
variables area under curve, g_accuracy,p_f measure and bit classification rate (BER) in addition to little bit
oversight rate.
Fig 3: Input Image Fig 3 a) Existing result Fig 3 b) Proposed result
LOAD IMAGE
APPLY SEGMENTATION
DETECT REGION OF INTEREST
APPLY FUZZY LOGIC BASED EDGE DETECTION
APPLY MODIFIED ADDITIVE HOUGH TRANSFORM
DETECT AND COLOR LANES
EVALUATE PARAMETERS

Fig 4: Input Image Fig 4 a) Existing result Fig 4 b) Proposed result
1.Area under curve:The area under a curve between two pointscanbefoundbydoingadefiniteintegralbetween
the two points. To find the area under the curve y = f(x) between x = a and x = b, integrate y = f(x) between the
limits of a and b. Areas under the x-axis will come out negative and areas above the x-axis will be positive.
Fig 5: Performance Analysis of Area Under Curve
2.Bit Classification Rate (BCR): Bit classification allows efficientselectionandclassificationcodesforeachbitare
generated by placing the bit style into the categorythatbestdescribesitsothatsimilarbittypesaregroupedwithin
a single category.
Fig 6: Performance Analysis of Bit Classification Rate
3. G_Accuracy: It is defined as number of instance per classes that have been correctly identified. Correctly
classified instances lead to the accuracy of results.

Fig 7: Performance Analysis of G_accuracy
It showing that the increase in theaccuracybyusinglaneimagesbasedonproposedadditiveHoughtransformover
the existing technique.
4 . P_F Measure: It is defined as the harmonic mean of sensitivity and specificity. Senstivity measures the
proportion of positives that are correctly identified as such (i.e. the percentage of sick people who are correctly
identified as having the condition). Specificity measures the proportionof negativesthatarecorrectlyidentifiedas
such (i.e., the percentage of healthy people who are correctly identified as not having the condition).
Fig 8: Performance Analysis of P F measure
5.conclusion
Lane detection enables us to obtain the position as well as direction of the vehicle along with lane information.
There are different types of methods that are used for detecting lines. The methods formulated until now are
operating effectively as well as providing beneficial results inscenario when the straight laneimagesaregenerally
there. However challenge is simply because that they are unsuccessful or otherwise not provide successful
outcomes whenever there are curved lane road images. In this modify Hough transform i.e. Fuzzy Logic is used to
improve straight lane as well as curved lane road images. The comparison has been drawn between Hough

transform and Fuzzy Logic by using various parameters area under curve, g_accuracy,p_f measure and bit
classification rate (BCR). The proposed technique has been designed and implemented in Matlab 2010a by using
image processingtoolbox.Theperformanceevaluationhasshowntheimprovementinproposedworkascompared
to the existing.
References
[1] Lin, Qing, Youngjoon Han, and Hernsoo Hahn. "Real-time lane departure detection based on extended edge-
linkingalgorithm"IEEESecondInternationalConferenceonComputerResearchandDevelopment,pp.725-730,May
2010.
[2] Kasai, Tatsuya, and Kazunori Onoguchi. "Lane Detection System forVehiclePlatooningusingMulti-information
Map”IEEE 13th International Conference on, Intelligent Transportation Systems (ITSC), pp. 1350-1356, Sep2010.
[3] Guo, Chunzhao, Seiichi Mita, and David McAllester. "Lane detection and tracking in challenging environments
based on a weighted graph and integrated cues" IEEE/RSJ International Conference on Intelligent Robots and
Systems (IROS),pp. 5543-5550, Oct2010.
[4] He, Jia, HuiRong, Jinfeng Gong, and Wei Huang. “A lane detection method for lane departure warning system"
IEEE International Conference on Optoelectronics and Image Processing (ICOIP), vol. 1, pp. 28-31, Nov2010.
[5] Leng, Yu-Chi, and Chieh-Li Chen. “Vision-based lane departure detection system in urban traffic scenes" IEEE
11th International Conference on control Automation Robotics & Vision (ICARCV), pp. 1875-1880, Dec 2010.
[6] Chen, Yong, Mingyi He, and Yifan Zhang. "Robust lane detection based on gradient direction" IEEE 6th
Conference onIndustrial Electronics and Applications (ICIEA),pp. 1547-1552, June 2011.
[7]Saeedi Xiaoyun, Wang, Wang Yongzhong, and Wen Chenglin. "Robust lane detection based on gradient-pairs
constraint" IEEE 30th Chinese Control Conference (CCC),pp. 3181-3185, July2011.
[8] Li, Jian, XiangjingAn, and Hangen He."Lane Detection Based on Visual Attention"IEEE Sixth International
Conference on Image and Graphics (ICIG),pp. 570-575, Aug 2011.
[9] Keyou, Guo, Li Na, and Zhang Mo. "Lane detection based on the random sample consensus" IEEEInternational
Conference on Information Technology, Computer Engineering and Management Sciences (ICM), vol. 3,pp. 38-
41,Sep 2011.
[10] Li, Hao, and FawziNashashibi. "Robust real-time lane detection based on lane mark segment features and
general a priori knowledge" IEEE International Conference on Robotics and Biomimetics (ROBIO), pp. 812-817,
Dec2011.
[11] Kang Deusch, Hendrik, Jürgen Wiest, StephanReuter,MagdalenaSzczot,MarcusKonrad,andKlausDietmayer.
"A random finite set approach to multiple lane detection" IEEE15th International Conference on Intelligent
Transportation Systems (ITSC),pp. 270-275, Sep2012.
[12] Singh, Rajandeep, and Prabhdeep Singh. "Integrated Lane Colorization Using Hough Transformation and
Bilateral Filter."International Journal of Engineering Sciences & Research Technology,Oct 2013, pp.
[13] Cassisi, Carmelo, Alfredo Ferro, Rosalba Giugno, Giuseppe Pigola, and Alfredo Pulvirenti."Enhancingdensity-
based clustering: Parameter reduction and outlier detection." Information Systems 38, no. 3 (2013): 317-330.
[14] Wu, Pei-Chen, Chin-Yu Chang, and Chang Hong Lin. "Lane-mark extraction for automobiles under complex
conditions." Pattern Recognition 47, no. 8 (2014): 2756-2767.
[15] Niu, Jianwei, Jie Lu, Mingliang Xu, Pei Lv, and Xiaoke Zhao. "Robust Lane Detection using Two-stage Feature
Extraction with Curve Fitting." Pattern Recognition 59 (2016): 225-233.
[16] Lin, Liang, Xiaolong Wang, Wei Yang, and Jian-Huang Lai. "Discriminatively trained and-or graph models for
object shape detection." IEEE Transactions on pattern analysisandmachineintelligence 37,no.5(2015):959-972.
[17] Niu, Jianwei, Jie Lu, Mingliang Xu, Pei Lv, and Xiaoke Zhao. "Robust Lane Detection using Two-stage Feature
Extraction with Curve Fitting." Pattern Recognition 59 (2016): 225-233.
[18] Jung, Soonhong, Junsic Youn, and Sanghoon Sull. "Efficient lane detection based on spatiotemporal
images." IEEE Transactions on Intelligent Transportation Systems 17, no. 1 (2016): 289-295.

Improved Performance of Fuzzy Logic Algorithm for Lane Detection Images

Recommended

More Related Content

What's hot (19)

Similar to Improved Performance of Fuzzy Logic Algorithm for Lane Detection Images (20)

More from IRJET Journal (20)

Recently uploaded (20)

Improved Performance of Fuzzy Logic Algorithm for Lane Detection Images