B04410814

IOSR Journal of Engineering (IOSRJEN) www.iosrjen.org
ISSN (e): 2250-3021, ISSN (p): 2278-8719
Vol. 04, Issue 04 (April. 2014), ||V1|| PP 08-14
International organization of Scientific Research 8 | P a g e
An Overview of Image processing for traffic applications
Nikita Sankhe, Poonam Sonar , Deven Patel
1
(M.E Student, Department of Electronics & Telecommunication, Rajiv Gandhi Institute of Tech.,
Mumbai University, India)
2
(Asst.Professor, Department of Electronics & Telecommunication, Rajiv Gandhi Institute of Tech.,
3(Asst.Professor, Department of Electronics & Telecommunication, Rajiv Gandhi Institute of Tech.,
Abstract: - The frequent traffic jams at major junctions call for an efficient traffic management system in place.
The resulting wastage of time and increase in pollution levels can be eliminated on a city-wide scale by these
systems. In this paper we present an overview of image processing and analysis tools used in these applications
and we relate these tools with complete systems developed for specific traffic applications. More specifically,
we categorize processing methods based on their input data as feature-driven, area-driven, or model-based and
the domain of processing as spatial or temporal frames.The image sequences from a camera are analyzed using
various edge detection and object counting methods to obtain the most efficient technique. Subsequently, the
number of vehicles at the intersection is evaluated and traffic is efficiently managed. The paper also proposes to
implement a real-time emergency vehicle detection system. In case an emergency vehicle is detected, the lane is
given priority over all the others.
Keywords: - Traffic monitoring, automatic lane finding, Object detection, Emergency vehicles, Image
Processing.
I. INTRODUCTION
The application of image processing and computer vision techniques to the analysis of video sequences
of traffic flow offers considerable improvements over the existing methods of traffic data collection and road
traffic monitoring. Existing methods include the detectors such as loop, radar, infrared, ultrasonic, and
microwave detectors which are expensive with limited capacity and involve installation, maintenance, and
implementation difficulties. Image processing offer a relatively low installation cost with little traffic disruption
during maintenance. Also they provide wide area monitoring allowing analysis of traffic flows and turning
movements, speed measurement, multiple-point vehicle counts, vehicle classification and highway state
assessment (e.g. congestion or incident detection).
Image processing also finds extensive applications in the related field of autonomous vehicle guidance,
mainly for determining the vehicle’s relative position in the lane and for obstacle detection. In section2 traffic
analysis can be done using automatic lane finding. In section 3 object detection discussed in detail and in section
4 emergency vehicle detection system is explained.
II. AUTOMATIC LANE FINDING
Automatic lane finding (ALF) is an important task for an adaptive traffic monitoring system. ALF can
assist and simplify the installation of a detection system. It enables applications in active vision systems, where
the camera viewing angle and the focal length of the camera lens may be controlled by the system operator to
find an optimum view. Based on the two features namely lane marking and road edges, lane finding is
characterized into two classes [5]
2.1 LANE REGION DETECTION
This class relates the detection of the lane with the changing intensity distribution along the region of a
lane. This class considers just the changes in the gray-scale values within an image sequence, without
addressing the problem of motion estimation which causes classification problem which labels image pixels into
road and non-road classes based on particular features. A typical classification problem involves the steps of
feature extraction, feature de-correlation and reduction, clustering and segmentation.

2.1.1 COLOR BASED SEGMENTATION
In this case, the features are defined by the spectral response of the illumination at the red, green and blue
bands. At each pixel, the (R, G, B) value defines the feature vector and the classification can be performed
directly on the (R, G, B) scatter diagram of the image.
Fig (1): color based segmentation [3, 5]
2.1.2 TEXTURE BASED SEGMENTATION
In this case, the texture of the image has been used as a feature for classification. The texture of the
road is normally smoother than that of the environment, allowing for region separation in its feature space. The
texture calculation can be based on the amplitude of the gradient operator at each image area which uses a
normalized gradient measure based on a high-resolution and a low-resolution (smoothed) image in order to
handle shadow interior and boundaries. [5]
Fig(2): texture based segmentation[5]
2.2 FEATURE-DRIVEN APPROACH
This class of approach is based on the detection of edges in the image and the organization of edges into
meaningful structures (lanes or lane markings). This class involves two levels of processing like feature
detection and feature aggregation.
2.2.1 FEATURE DETECTION
This part aims at extracting intensity discontinuities. To make the detection more effective, image
enhancement is performed followed by a gradient operator. The dominant edges are extracted based on
thresholding of the gradient magnitude. At this stage, the direction of edges at each pixel can be computed based
on the phase of the gradient and a curvature of line segments can be estimated based on neighborhood relations.
[6]

2.2.2 FEATURE AGGREGATION
This process organizes edge segments into meaningful structures (lane markings) based on short range
or long-range attributes of the lane. Short-range aggregation considers local lane fitting into the edge structure of
the image. Long-range aggregation is based on a line intersection model, based on the assumption of smooth
road curvature. Thus gross road boundaries and markings must be directed towards a specific point in the
image, the focus of expansion (FOE) of the camera system. Along these directions, it detects lane markings
through a horizontal (linear) edge detector and enhances vertical edges via a morphological operator. For each
horizontal line, it then forms correspondences of edge points to a two-lane road model (three lane markings) and
identifies the most frequent lane width along the image, through a histogram analysis.
In general, feature driven approaches are highly dependent on the methods used to extract features
and they suffer from noise effects and irrelevant feature structures. Often in practice the strongest edges are not
the road edges, so that the detected edges do not necessarily fit a straight-line or a smoothly varying model.
Shadow edges can appear quite strong, highly affecting the line tracking approach.[3,6]
III. OBJECT DETECTION
Some fundamental issues of object detection are considered in this section. Different approaches have been
categorized according to the method used to isolate the object from the background on a single frame or a
sequence of frames.
3.1 THRESHOLDING
This is one of the simplest, but less effective techniques, which operates on still images. It is based on the
notion that vehicles are compact objects having different intensity form their background. Thus by thresholding
intensities in small regions we can separate the vehicle from the background. This approach depends heavily on
the threshold used, which must be selected appropriately for a certain vehicle and its background. Adaptive
thresholding can be used to account for lighting changes, but cannot avoid the false detection of shadows or
missed detection of parts of the vehicle with similar intensities as its environment. To aid the thresholding
process, binary mathematical morphology can be used to aggregate close pixels into a unified object.
Furthermore, gray-scale morphological operators have been proposed for object detection and identification that
are insensitive to lighting variation.[1,3,5]
Fig(3): Object detection done using thresholding[3]
1.1.1 MULTI GRID IDENTIFICATION OF REGIONS OF INTEREST
This method first generates a hierarchy of images at different resolutions. Subsequently, a region search
begins at the top level (coarse to fine). Compact objects that differ from their background remain distinguishable
in the low-resolution image, whereas noise and small intensity variations tend to disappear at this level. Thus,
the low-resolution image can immediately direct attention to the pixels that correspond to such objects in the
initial image. Each pixel of interest is selected according to some interest function which may be a function of
the intensity values of its adjacent pixels, edge strength, or successive frame differencing for motion analysis.
[4]
3.2 EDGE-BASED DETECTION (SPATIAL DIFFERENTIATION)
Approaches in this class are based on the edge-features of objects. They can be applied to single images
to detect the edge structure of even still vehicles. Morphological edge-detection schemes have been extensively
applied, since they exhibit superior performance. In traffic scenes, the results of an edge detector generally

highlight vehicles as complex groups of edges, whereas road areas yield relatively low edge content. Thus the
presence of vehicles may be detected by the edge complexity within the road area, which can be quantified
through analysis of the histogram. Alternatively, the edges can be grouped together to form the vehicle’s
boundary.[2,6]
Fig(4): object detection done using different edge detection algorithm[6]
Fig.(4.a) input image is given. In Fig (4.b), Boolean edge detector is used to perform a decent job of
marking the locations of edges; however it failed to complete the edges making object detection difficult. In Fig
(4.c), Marr-Hildreth detector is applied since it is susceptible to noise and gave a lot of false results. In Fig (4.d)
and (4.e), Sobel and Prewitt operators are used for recognizing edges that are horizontal or vertical and are
susceptible to noise. And Fig,(4.f) Canny edge detector is used which basically uses two thresholds which used
for detecting true weak edges, providing a better and fairly noise resistant output.[6]
3.3 BACKGROUND FRAME DIFFERENCING
This method is based on forming a precise background image and using it for separating moving objects
from their background. The background image is specified either manually, by taking an image without
vehicles, or is detected in real-time by forming a mathematical or exponential average of successive images. The
detection is then achieved by means of subtracting the reference image from the current image. Thresholding is
performed in order to obtain presence/absence information of an object in motion. The background can change
significantly with shadows cast by buildings and clouds, or simply due to changes in lighting conditions. With
these changing environmental conditions, the background frame is required to be updated regularly. There are
several background updating techniques. The most commonly used are averaging and selective updating. In
averaging, the background is built gradually by taking the average of the previous background with the current
frame. In selective updating, the background is replaced by the current frame only at regions with no motion
detected; where the difference between the current and the previous frames is smaller than a threshold. Selective
updating can be performed in a more robust averaging form, where the stationary regions of the background are
replaced by the average of the current frame and the previous background. [4, 6]
(a) (b) (c)
(d) (e) (f)

(a) Real-time image [6] (b) Background image[6]
(c) Subtracted image[6] (d) No. of vehicles = 3[6]
II. EMERGENCY VEHICLE DETECTION
Fig(5): Emergency vehicle detection[7]
The various steps of detection system are shown in following flow-chart. In case of normal regulation the
image is subjected to adaptive background subtraction, edge detection, and object counting methods. Using this,
the number of vehicles is calculated and traffic is regulated as per the congestion in each lane. [4,7]

2.1 FLOWCHART FOR EMERGENCY VEHICLE DETECTION
In case a red beacon is detected, the next step is to identify whether it is from an emergency vehicle or not.
This is done by identifying the blinking frequency of red light detected in the image sequence and comparing it
to the standard used by the emergency vehicles as shown in the above flow chart. This system will override the
normal regulation system and will change light to green till emergency vehicle is exited. Then system wiil start
running for normal regulation.
IV. CONCLUSION
In this paper, we have studied different parameters which are used in traffic scene analysis. Automatic
lane detection based on region and texture .In object detection we have studied Thersholding, Background
Subtraction, Boolean Edge Detector, Sobel & Prewitt operator and Canny Edge Detector . Of all studied
methods Canny Edge Detector is gives error free output even in noise and for weak signals .Thus Emergency
vehicle Detection can be well implementated using Canny Edge Detector which tracks the red beacon signal for
more accurate detection.
V. ACKNOWLEDGEMENT
The author would like to thank the following for valuable discussions related to this work: Prof. Poonam Sonar,
Prof. S.D.Patil and Prof. Deven Patel.
VI. REFERENCES
[1] V. Kastrinaki, M. Zervakis, K. Kalaitzakis ”A survey of video processing techniques for traffic
applications” Image and Vision Computing 21 (2003) 359–381. K.Compton, S.Hauck,
[2] Benjamin Coifmana, David Beymerb, Philip McLauchlanb, Jitendra Malikb”A real-time computer vision
system for vehicle tracking and traffic surveillance” Journal of Artificial Intelligence Research 3 (1995)
325–348.Volume 6, Issue 4, August 1998, Pages 271–288..
[3] B.D. Stewart, I. Reading, M.S. Thomson, T.D. Binnie, K.W.Dickinson, C.L. Wan, Adaptive lane finding
in road traffic image analysis, Proceedings of Seventh International Conference on Road Traffic
Monitoring and Control, IEE, London (2003).
Fig (6): Flowchart for emergency vehicle detection [7]

[4] Vikramaditya dangi by, Amol Parab, Kshitij Pawar & S.S Rathod “Image processing based intelligent
traffic controller”, Undergraduate Academic Research Journal (UARJ), ISSN : 2278 – 1129, Volume-1,
Issue-2, 2012.
[5] M. Fathy and M. Y. Siyal “A window-based image processing technique for quantitative and qualitative
analysis of road traffic parameters”
[6] M. Xie, L. Trassoudaine, J. Alizon, J. Gallice, Road obstacle detection and tracking by an active and
intelligent sensing strategy, Machine Vision and Applications 7 (1994) 165–177.
[7] A. Broggi, S. Berte, Vision-based road detection in automotive systems: a real-time expectation-driven
approach, Journal of Artificial Intelligence Research 3 (1995) 325–348.
[8] M. Bertozzi, A. Broggi, GOLD: a parallel real-time stereo vision system for generic obstacle and lane
detection, IEEE Transaction Image Processing 7 (1) (1998).
[9] Ahmed S. Salama, Bahaa K. Saleh, Mohamad M. Eassa, ”Intelligent Cross Road Traffic Management
System (ICRTMS),” 2nd Int. Conf. on Computer Technology and Development, Cairo, Nov 2010, pp.
27-31.
[10] Y. Wu, F. Lian, and T. Chang, “Traffic monitoring and vehicle tracking using roadside camera,” IEEE
Int. Conf. on Robotics and Automation, Taipei, Oct 2006, pp. 4631– 4636
[11] K. Wang, Z. Li, Q. Yao, W. Huang, and F. Wang, “An automated vehicle counting system for traffic
surveillance,” IEEE Int.Conf. on Vehicular Electronics and Safety, Japan, Dec 2007, pp. 1-6.
[12] K. Hang, “Real-time image acquisition and processing system design based on DSP,” The 2nd IEEE Int.
Conf. on Computer and Automation Engineering (ICCAE), Singapore, Feb 2010, pp. 492-496.
[13] B. Fazenda, H. Atmoko, F. Gu, L. Guan1 and A. Ball, ”Acoustic Based Safety Emergency Vehicle
Detection for Intelligent Transport Systems,” ICCAS-SICE, Fukuoka, Aug 2009, pp.4250-4255.

B04410814

More Related Content

What's hot (17)

Similar to B04410814 (20)

More from IOSR-JEN (20)

Recently uploaded (20)

B04410814