Design and development of image processing algorithms for quantitative road traffic data analysis

• Main Author: Krishnalal Puthiya Veettil.
• Other Authors: Mohammed Yakoob Siyal
• Format: Final Year Project
• Language: English
• Published: 2013
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/53397
• Institution: Nanyang Technological University

High traffic volume monitoring and traffic congestion being a major concern for authorities, this project targets to design quantitative analysis algorithms based on image processing technique. Image processing techniques in MATLAB software were used to realize the requirements. Several other existing traffic data collection and monitoring techniques were studied prior to implementation of this project that included advantages and disadvantages of each technique. Traffic video footage samples were collected from express highways in both night and day conditions and extracted into frames before going through image segmentation techniques such as edge detection and background difference. Algorithms to obtain quantitative information such as vehicle count and vehicle speed were designed which incorporated to MATLAB Graphic User Interface (GUI) environment in which the results of processing were displayed with a number of user defined settings. Vehicle classification was also done based on its size. Further applications such as monitoring of heavy traffic were also developed. Results comparison between different segmentation methods was performed to check the best detection method. Edge detection techniques generally showed comparatively better results than other methods. Certain considerations that may affect the performance of the project and Recommendations for further improvements are discussed at the end of report. In a nutshell, this project provided the author an insight about numerous traffic data analysis techniques, algorithm development and data analysis.