The document presents a vision-based traffic surveillance system that uses digital image processing techniques. The system works to improve image quality by enhancing contrast and removing noise and blurring. It then uses edge detection and morphological processing to segment vehicles. The number of vehicles in each lane is counted and used to determine the time allotted for that lane, with accuracy of 90% compared to existing systems.

1. Vision Based Traffic Surveillance System Presented By S.maheswaran

2. Improvement of Pictorial Information To improve the contrast of the image To remove noise To remove blurring caused by movement of the image acquisition devices To correct geometrical distortions caused by the lens Why Digital Image Processing Automatic Machine perception for intelligent interpretation of scenes or pictures.

3. Different Fields of Applications Character Recognition & Signature Verification Industrial Process Monitoring Biometrics and Forensic ( Recognition and Verification of persons using Face, Palm & Fingerprints ) Military surveillance and Target Identification Remote Sensing ( Satellite Image Processing) Safety and Security ( Night vision) Biomedical Engineering ( Diagnosis and Surgery ) Traffic monitoring

4. To monitor the road To initiate automated vehicle tracking Lane Jam Detection To measure the number of vehicles in the lane To recognize number plates of the vehicles Need for Vision Based Traffic Control

5. Image Acquisition Preprocessing Feature Extraction Morphological Processing Decision Methodology Histogram Equalization Median Filtering

6. Image Acquisition Initial Image Acquired from traffic lane Original Image Gray Scale Image

7. Histogram - plot between p(r k ) vs r k To enhance the contrast of the traffic image Histogram Equalization Defined as mapping of gray levels p into gray levels q such that the distribution of gray level q is uniform

8. Output of Histogram Equalization Original Image Enhanced Image

9. To reduce the noise in image, Median filtering is used. Provide excellent noise reduction. Median is calculated by first sorting all the pixel values from the surrounding neighborhood into numerical order and then replacing the pixel with middle pixel value. Filtering

10. Output of 3x3 Median Filtering Enhanced image After 3X3 median filtering

11. Segmentation - Subdivides an image into its constituent regions or objects. Edge - A set of connected pixels whose two-dimensional first order derivative is greater than a specified threshold. Edge Detection - 2-D gradient based approach is used. (Sobel Mask) Segmentation

12. Edge Detection Gradient G x in X-direction G x = (z 7 +2z 8 +z 9 ) – (z 1 +2z 2 +z 3 ) Gradient G y in Y direction G y = (z 3 +2z 6 +z 9 ) – (z 1 +2z 4 +z 7 ) Mask-1 Mask-2 -1 -2 -1 0 0 0 1 2 1 -1 0 1 -2 0 2 -1 0 1

13. Gradient of an image f(x,y) at location (x,y) is defined as the vector M = [G 2 x + G 2 y] ½ The direction of the gradient vector of an image f(x, y) at location (x,y) is given as α (x,y) = tan -1 [Gy / Gx] Cont..

14. Median filtered Image Edges of vehicle Output of Edge Detection

15. It deals with tools for extracting image components that are useful in the representation and description of region shape. It has two main process, 1. Dilation 2. Region Filling Morphological Processing

16. Dilation This process helps to thicken the edges. The output of dilation results in the foreground pixels are represented by 1's and background pixels by 0's. 3×3 square structuring element is taken for our work with the origin at its center. 1 1 1 1 1 1 1 1 1

17. If an image denotes a subset containing pixels, whose elements are 8-connected boundary points of region beginning with a point p inside a boundary, the objective is to fill the entire image with 1's. Dilated Image After Region Filtering Region Filling

18. Decision The vehicles are counted after back ground elimination using certain algorithm. 1’s refers to vehicles and 0’s refers to free space. By counting number of 1’s we can find the number of vehicles in the lane. Number of Vehicles is 5 . So the time allotted for that lane is 30 sec. 120 sec (default time) > 95 5 100 sec 75 - 95 4 60 sec 50 - 75 3 45 sec 25 -50 2 30 sec < 25 1 Time Allotted No of Vehicles S.No

19. Conclusion 20 lane image were taken for analysis. 90% accuracy was obtained. Time allotment to the channel is appropriate & better than existing µ c and sensor based system. Recognition of number plates and speed of the vehicles is possible.

20. References E. Atko¡ci unas1, R. Blake, A.Juozapavi¡cius, M. Kazimianec, Nonlinear Analysis: Modelling and Control , 2005, Vol. 10, No. 4, 315–332 Image Processing in Road Traffic Analysis. 2. G.D. Sullivan, K. Baker, et al. Model-based Vehicle Detection and Classification using Orthographic Approximations , in: Proc. British Machine Vision AssociationConference, 1996. 3. D.A. Forsyth, J. Ponce. Computer Vision. A Modern Approach , Prentice Hall, 2003.

21. 4. D. Beymer, et al. Computer Vision System Measuring Traffic Parameters , in: Proc. IEEE Conf. On Computer Vision and Pattern Recognition, 1977. 5. L.G. Shapiro, G.C. Stockman. Computer Vision , Prentice Hall, 2001. 6. L.A. B. Jähne, H. Haußecker, P. Geißler. Computer Vision and Applications , Academic Press, 1999. Cont..

22. Open For Discussion

23. Thank you