Evaluating The Performance of Integrated Lane Colorization Using Hough Transformation and Bilateral Filter

International Journal of Emerging Trends & Technology in Computer Science (IJETTCS)
Web Site: www.ijettcs.org Email: [email protected], [email protected] Volume 3, Issue 1, January February 2014 ISSN 2278-6856
EVALUATING THE PERFORMANCE OF INTEGRATED LANE COLORIZATION USING HOUGH TRANSFORMATION AND BILATERAL FILTER
Rajandeep Singh1, Prabhdeep Singh2
2
Department of Computer Science and Engineering, GIMET, Amritsar Assistant Professor, Department of Computer Science and Engineering, GIMET, Amritsar
Abstract: Lane coloration is a significant technique in a

number of intelligent automobile applications, containing the lane trip recognition and warning board, intelligent journey control and autonomous driving. This paper presents a literature review on the techniques for lane coloration and explores the benefits and limits of existing lane colorization problems. It has been found that most of existing researchers has neglected the filtering and restoration techniques. However it is also found that existing researchers has also neglected the overheads of existing techniques. So in order to reduce the limitations of the existing researchers we have proposed a new strategy which uses bilateral filter as preprocessing stage which has ability to reduce the noise from images before further processing. By doing so it has started working fine even for noisy images. The proposed algorithm has been designed and implemented in MATLAB. By passing different images we have shown the significant improvement of the proposed algorithm over the existing algorithms.
tracking processes to put properly strong constraints on the likely location and orientation of the lane edges in a new image. Lane coloration technique has to locate the lane edges without any prior knowledge of the road geometry, and do so in situations where there may be a countless clutter in the road image. This clutter can be because of the noise, dust, shadows, puddles, oil stains, tire skid marks, etc. Thus it becomes a major issue when noise is present in the input image. Thus we focus on providing an efficient algorithm which will provide better results when noise or any other unknown factor is present in the image.
2. LITERATURE SURVEY
Christian et al. (2005) has presented a multi-camera lane coloration algorithm that makes use of a conventional PC and a graphics card. The feature detection and lane fitting approach are able to cope with different lighting situations, weather conditions, road layouts and lane markings. They also concluded that the lane colorization is an important application in intelligent vehicular systems. Tseng et al. (2005) has discussed a lane marking detection algorithm by using geometry information and modified Hough transform. First, the acquired image is divided into road part and non-road part from the geometry information. Secondly, the histogram of intensities is applied to quantize the road image into a binary image. Thirdly, a modified Hough transform method is developed to detect the lane markings in road image by using the road geometry information. However, only straight lane marking is studied here in [2]. Thus, it is interesting to develop the detection algorithm for curved lane markings. Shanshan et al. (2012) has used the Canny edge detection, improved Hough transform for the detection of straight lines, through set a proper threshold to get rid of the false road edge, fit out the actual road edge. The experimental results show that the proposed algorithm has strong ability to adapt to the environment, and the extraction of boundary has a high precision. How to deal Page 152
Keywords: Image filtering, Lane coloration, Overheads, Hough transformation, bilateral filter.
1. INTRODUCTION
Lane coloration plays a significant role in a number of intelligent automobile applications. Various lane coloration methodologies have been proposed so far by different researchers. They are classified into infrastructure-based and vision-based approaches. Although the infrastructure-based methods accomplish extremely robustness, building cost to lay leaky coaxial cables or to put magnetic indicators on the road surface is high. Vision based methods with camera on a vehicle have benefits to use well-known available lane detection in the road location and to sense a road curvature in front view. Lane coloration and lane tracking are two different steps in vision based techniques. Lane coloration is the problem of discovering lane boundaries without any prior information of the road. Lane tracking deal with the tracking of the lane edges from frame to frame given an existing model of road geometry. Lane tracking is quite simple problem than lane coloration, as knowledge of the road geometry is known in advance which permits lane Volume 3, Issue 1 January February 2014

with the road edge detection under the complex environment still needs to do further research. Miao et al. (2012) has proposed a novel intelligent vehicle oriented lane coloration approach. Conclusions are made as (1) .A five steps lane coloration scheme that can successfully locate the lane line or boundary. In addition, it is also effective in various bad road scenes. (2) No assumptions are made about road structure, marking, or lane type, etc, so it owns a better generalize capability than others. (3) Plenty of experiments have been conducted and results show that the proposed method is robust to noises, shadows, illumination variations in the captured road videos, and is also applicable to both the marked and the unmarked road. Zaidi et al. (2012) has also suggested that the bottleneck in traffic is sometimes due to bus stops that are present on highways. Since we at the movement have very little information that the bus stops might have on certain highways, Zaidi et al. (2012) usually were not able to plan their locations well enough with respect to lane congestions. If only we have precise information of traffic concentration on each lane we can plan bus stops, route traffic and ovoid congestions. Hongying et al. (2013) has described lane coloration and tracking method based on annealed particle filter algorithm, which combines multiple cues with annealed particle filter. As a first step, pre-processing, with bar filter and color cues being used. In the annealed particle filter step, angle information of edge map is utilized to measure weights of particles. Saha et al. (2012) has presented an algorithm for detecting marks of road lane and road boundary with a view to the smart navigation of intelligent vehicles. Initially, it converts the RGB road scene image into grey image and employs the flood-fill algorithm to label the connected components of that grey image. Afterwards, the largest connected component which is the road region is extracted from the labelled image using maximum width and no. of pixels. Dhana Lakshmi et al. (2012) has exposed that lane coloration is an essential component of Advanced Driver Assistance System. The cognition on the roads is increasing day by day due to increase in the four wheelers on the road. The cognition coupled with ignorance towards road rules is contributing to road accidents. The lane marking violence is one of the major causes for accidents on highways in India. Cuong et al. (2012) has proposed a method for robust detection of pedestrian marked lanes at traffic crossings. The proposed method employed colour and intensity information in extracting the candidate markers and verified the extracted markers in a probabilistic framework. Multiple geometric cues were used for the verification.
3. HOW LANE COLORIZATION WORKS

Following are the different steps of the lane colorization using bilateral filter are given. Each steps convert and image into a new form. Like bilateral filter will produce noise free image, global histogram will reduce the unwanted objects etc. Step 1: Read the Road image Step 2: Apply bilateral filter to remove the noise from the input image so that image is more suitable to the rest of the application. Step 3: Convert image into the gray scale if it is in colour plane. Step 4: Now global histogram based thresholding will be applied to detect the background. Step 5: Now detected background will be removed or subtracted from the main road image. Step 6: Now canny edge detection will be applied to detect lane edges. Step 7: Now apply Hough transform and thresholding to segment the image and detect the lanes. Step 8: Now just color the lanes
Figure 1: How to capture lanes Figure 1 is demonstrating the vehicles as nodes, the main vehicles sensor or camera will capture the road at a time and based upon the captured image lane detection will detect the lanes generally called as subject lane (lane of interest) and additional lanes.
4. EXPERIMENTAL RESULTS
By taking different road images for experimental purpose we have seen the results of the integrated and existing approach. It is shown in the following figures why proposed algorithm is more beneficial over existing in case of noisy images. Figure 2 is showing the noisy input image. It is clearly shown that the visibility of the image is quite poor.
Figure 2: Input noisy image Volume 3, Issue 1 January February 2014 Page 153

is quite improved than the image shown in figure 2. Figure 4 has demonstrated the output of the Hough transformed without using the bilateral filter. It has been noticeably shown that the Hough lines are not as accurate as expected. Figure 5 has shown the Hough transformed output image using bilateral filter. The results are quite better than the image shown in figure 4. The lane colorized image is shown in Figure 6 and 7. The image shown in figure 6 is without bilateral filter so have some artifacts i.e. not visibility too accurate and even lanes are not properly detected. But image shown in Figure 7 is showing the smoothed image even the colorized lanes are properly shown. Thus proposed algorithm is quite better than the existing algorithm.
Figure 3: Filtered image using bilateral filter
Figure 4: Hough transformed output without using bilateral filter Figure 7: Final output image t using bilateral filter
5. PERFORMANCE ANALYSIS
This section contains the performance comparison of the proposed algorithm and existing algorithms by taking different performance parameters. The overall objective of this chapter is to prove that the proposed algorithms provide more accurate results than the existing algorithms. Table 1: Accuracy Analysis Image Old method Proposed Technique 1 94.33 99.93 2 94.41 99.97 3 97.31 99.26 4 96.65 99.69 5 98.14 99.77 6 93.12 98.76 7 96.5 98.91 8 97.56 99.49 9 91.33 99.37 10 94.66 99.66 11 95.6 98.74 12 93.6 97.97 13 98.2 99.49 14 96.6 99.34 15 92.12 99.73 Page 154
Figure 5: Hough transformed output using bilateral filter
Figure 6: Final output without using bilateral filter Figure 3 is showing the filtered image using bilateral filter. It is clearly shown that the visibility of the Figure 3 Volume 3, Issue 1 January February 2014

Table 1 is showing the accuracy analysis of the proposed and exiting technique. It is found that the accuracy of the proposed algorithm in case of the input image has shown quite effective results than the existing method. Accuracy is need to as much as possible. The accuracy of the proposed technique is more than 98.74 in the most of cases therefore the proposed algorithm is quite accurate than the others. Table 2: Bit Error Rate (BER) Evaluation Image Old method Proposed Technique 1 44.33 5.93 2 34.41 0.97 3 27.31 0.26 4 26.65 0.69 5 18.14 9.77 6 3.12 1.76 7 6.5 2.91 8 7.56 3.49 9 1.33 1.37 10 4.66 0.66 11 5.6 2.74 12 3.6 1.97 13 18.2 0.49 14 16.6 0.34 15 12.12 0.73 Table 2 has shown the BER investigation of the proposed and exiting procedure. It is found that the BER of the proposed procedure in case of the input images shown in has given fairly effective outcomes than the existing technique. As required BER need to be reduced. It is clearly shown that BER is quite less in proposed algorithm reason behind this is the O (1) bilateral filter. Table 3: Specificity Evaluation Old method Proposed Technique .94 . 99 .64 .98 .31 .98 .65 .98 .14 .98 .12 .89 .95 .99 .56 .98 .83 .98 .86 .89 .76 .83 .86 .95 .82 .91 .86 .93 .82 .92 Table 3 has shown the Specificity exploration of the proposed and available technique. As specificity needs to be maximized therefore it is proved that the Specificity of the proposed technique in case of the input images has given objectively effective results than the surviving technique. It is clearly shown that in many cases we have achieved specificity up to .99 which is almost equal to 1. Therefore we can justify in terms of specificity that the proposed algorithm is quite effective and giving accurate results. Table 4: PSNR Evaluation Image Old method Proposed Technique 1 9.99 28.98 2 10.32 32.83 3 13.33 24.79 4 14.52 30.42 5 16.14 29.34 6 11.69 28.98 7 9.89 24.48 8 9.57 25.65 9 11. 23 33.54 10 13.11 32.24 11 12.2 29.29 12 13.58 31.28 13 17.89 34.27 14 14.56 33.38 15 13.87 37.39 Table 4 has shown the PSNR examination of the planned and traditional method. It is proved that the PSNR of the proposed technique in case of the input images has specified quantitatively improved consequences than the persisting technique.
Image 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Figure 8: Accuracy Analysis Figure 8 has shown the accuracy analysis of the proposed and exiting technique. Magenta color shows the result of old technique whereas new technique is shown by using blue color. Y axis has shown the accuracy rate or Page 155
Volume 3, Issue 1 January February 2014

percentage. X axis has shown the input set of images from 1 to 15. Figure 8 has demonstrated that the accuracy of the proposed algorithm in case of the input images has shown quite effective results than the existing method. Accuracy is need to as much as possible. The accuracy of the proposed technique is more than 98.74 in the most of cases therefore the proposed algorithm is quite accurate than the others. proposed algorithm is quite effective and giving accurate results.
Figure 11: PSNR comparison Figure 11 has shown the PSNR examination of the planned and traditional method. It is proved that the PSNR of the proposed technique in case of the input images has specified quantitatively improved consequences than the persisting technique.
Figure 9: Bite Error Rate (BER) Analysis Figure 9 has publicized the BER exploration of the projected and exiting procedure. It is established that the BER of the proposed technique in case of the input images has specified objectively effective outcomes than the existing technique. As required; BER need to be reduced. It is clearly shown that BER is quite less in proposed algorithm reason behind this is the O (1) bilateral filter. Therefore the proposed algorithm has shown quite effective results in case of the BER.
6. CONCLUSION
Lane coloration is becoming popular in real time vehicular ad-hoc network. The methods developed so far are working efficiently and giving good results in case when noise is not present in the images. But problem is that they fail or not give efficient results when there is any kind of noise in the road images. The noise can be anything like dust, shadows, puddles, oil stains, tire skid marks, etc. So in order to reduce these problems a new strategy is proposed which has integrated Hough transform based lane colorization with the bilateral filter. The integrated approach has shown significant improvements over the existing methods especially when noise is present in the images. The performance evaluation is also done by considering various wellknown image parameters. The parameters evaluation has also shown quite effective results.
7. FUTURE WORK
Figure 10: Specificity exploration analyses Figure 10 has shown the Specificity exploration of the proposed and available technique. As specificity needs to be maximized therefore it is proved that the Specificity of the proposed technique in case of the input images has given objectively effective results than the surviving technique. It is clearly shown that in many cases we have achieved specificity up to .99 which is almost equal to 1. Therefore we can justify in terms of specificity that the Volume 3, Issue 1 January February 2014 Due to the non-availability of the real time environment the simulation environment has been used to implement and verify the proposed algorithms accuracy. Therefore dissertation is simulation based. In near future we will use embed programming to validate the proposed work in efficient manner. However in this work; canny edge detection is used, but now many edge detectors has been developed so far which works more accurately and detect edge in more efficient manner. So in near future we will use fuzzy logic based edge detectors to improve the performance and accuracy of the proposed algorithm further. Page 156

Web Site: www.ijettcs.org Email: [email protected], [email protected] Volume 3, Issue 1, January February 2014 ISSN 2278-6856 REFERENCES
[1] Christian Lipski, Bjorn Scholz, Kai Berger, Christian Linz and Timo Stich, A Fast and Robust Approach to Lane Marking Detection and Lane Tracking, Computer Graphics Lab, pp. 1-4, 2005. [2] Chien-Cheng Tseng, Hsu-Yung Cheng and BorShenn Jeng, A Lane coloration Algorithm Using Geometry Information and Modified Hough Transform, 18th IPPR Conference on Computer Vision, Graphics and Image Processing, pp. 796-802, 2005. [3] Shanshan Xu, Jie Ying and Yanbin Song, Research on Road Detection Based on Blind Navigation Device, IEEE, pp. 69-71, 2005. [4] Xiaodong Miao, Shunming Li and Huan Shen, Onboard lane coloration system for intelligent vehicle based on monocular vision, International journal on smart sensing and intelligent systems, vol. 5, no. 4, pp. 957- 972, 2012. [5] Shoaib Zaidi, Mir Shabbar Ali, Sohaib Nomani, Annus Bin Khalid and Fawad Shamim, Automated lane coloration for vehicular traffic, NED University of Engineering and Technology, 2012. [6] Hongying Zhao, Zhu Teng, Hong-Hyun Kim, and Dong-Joong Kang, Annealed Particle Filter Algorithm Used for Lane coloration and Tracking, Journal of Automation and Control Engineering, Vol. 1, No. 1, January 2013. [7] Anik Saha, Dipanjan Das Roy, Tauhidul Alam and Kaushik Deb, Automated Road Lane coloration for Intelligent Vehicles, Global Journal of Computer Science and Technology Vol. 12, 2012. [8] M DhanaLakshmi and B.J. Rani Deepika, A brawny multicolor lane coloration method to indian scenarios, IJRET OCT, pp. 202 206, 2012. [9] Manh Cuong Le, Son Lam Phung and Abdesselam Bouzerdoum, Pedestrian Lane coloration for Assistive Navigation of Blind People, 21st International Conference on Pattern Recognition (ICPR 2012), pp 2594- 2597 2012. includes analysis of algorithms, computer networking, digital image processing, database management system and programming languages. He has recently started research work on lane detection and colorization.
AUTHOR
Rajandeep Singh Gill is a M.Tech student in Global College Amritsar (Pb.) India. His research interest includes image processing and vision applications. His research work on lane colorization has put great impact on the researchers. He has great knowledge of image processing toolbox in MATLAB.
Prabhdeep Singh is an assistant professor at Global College Amritsar (Pb.) India. His research interest includes image processing and computer networking. His expertise Volume 3, Issue 1 January February 2014 Page 157

Evaluating The Performance of Integrated Lane Colorization Using Hough Transformation and Bilateral Filter

Uploaded by

Copyright:

Available Formats

Evaluating The Performance of Integrated Lane Colorization Using Hough Transformation and Bilateral Filter

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Evaluating The Performance of Integrated Lane Colorization Using Hough Transformation and Bilateral Filter

Uploaded by

Copyright:

Available Formats

International Journal of Emerging Trends & Technology in Computer Science (IJETTCS)

Abstract: Lane coloration is a significant technique in a

International Journal of Emerging Trends & Technology in Computer Science (IJETTCS)

3. HOW LANE COLORIZATION WORKS

International Journal of Emerging Trends & Technology in Computer Science (IJETTCS)

Figure 3: Filtered image using bilateral filter

Figure 5: Hough transformed output using bilateral filter

International Journal of Emerging Trends & Technology in Computer Science (IJETTCS)

Volume 3, Issue 1 January February 2014

International Journal of Emerging Trends & Technology in Computer Science (IJETTCS)

International Journal of Emerging Trends & Technology in Computer Science (IJETTCS)

You might also like