Proceedings of The 4th International Conference on Advanced Research in Supply Chain Management
Year: 2024
DOI:
[PDF]
Digital Signal Process
Clean Chakma
ABSTRACT:
In intelligent transportation systems, the area is growing at a speedy rate, and the use of Digital Signal Processing (DSP) techniques for traffic light detection has emerged as a critical solution for the development of an autonomous drive and monitoring of traffic. This project demonstrates the method using MATLAB-based image processing for more accurate colour detection of traffic lights. The traffic infrastructure may be much more intelligent and effective. The combination of grayscale conversion, edge detection, and dilation are all tools that help in the ability to find the color of the traffic light and accordingly show the relevant signal. This is applicable in the real scenario; hence, this is an immense advancement in the application of DSP. This is considered a high-complexity engineering challenge, which requires a complex blend of knowledge in digital signal processing, good programming skills, and deep insight into design capabilities offered by MATLAB. Therefore, this project discusses addressing this problem with minimal resources but emphasizes minimal hardware and MATLAB software, though striving for a higher level of accuracy and efficiency in constrained processing memory and time. This asserts the lack of universality in the solution and necessarily involves admitting that there might be different hardware and software configurations. Diverse results in relation to costs, performance, and resource utilization are related to this. It also addresses common challenges in image processing, such as the variation in illumination, camera angles, and noise interference. This proposed solution will go in-depth into Convolutional Neural Networks (CNN), template matching techniques, and sensor-based systems with a hybrid approach involving machine learning and traditional image processing methods. The selected methodology will start from the time of image capturing through to its processing by applying edge detection and dilation techniques to get accurate detection of the traffic light colors. This project concludes by detecting and recognizing the color of the traffic light successfully with the application of edge detection, labeling, and color-thresholding techniques, in turn with the developed MATLAB code. On the other hand, it also shows some limitations in dealing with complicated traffic scenarios and variances in image conditions. Thus, practical applications of DSP techniques in the solution of problems related to transportation prove the outlined significant role of DSP on the way to improving the functionality and reliability of modern intelligent transportation systems.
keywords: Traffic Light Detection, Image Processing Techniques, MATLAB Programming, Intelligent Transportation Systems, Edge Detection and Dilation, Digital Signal Processing, Autonomous Transportation Systems, Convolutional Neural Networks (CNNs)