Comparison of Edge Detection Techniques and Mathematical Morphology in Car Plate Detection Application

License plate detection technology is used in a variety of applications, including parking and traffic monitoring systems. Furthermore, license plate detection is also an important application of image processing. Previously, there have been numerous studies based on different edge detection techniques for license plate recognition, involving the Roberts operator, Prewitt operator and Canny algorithm. However, no comprehensive comparisons between different techniques have been done. Besides, the impact of vehicle color and morphological parameters on edge detection capability was not discussed. In this article, five different edge detecting techniques are used to detect four vehicle images in different colors. The evaluation of edge detection capability is also performed by using the Mean-Square Error (MSE), Peak Signal-to-Noise Ratio (PSNR), and Time Consumed (TC) index. The effects of edge detection techniques and vehicle color on license plate detection are also compared. Additionally, the morphological operation with different window sizes for the plate extraction is also performed. The observations demonstrate that vehicle color has a significant impact on the edge detection of vehicle plate, while the performance of different techniques for the same vehicle is negligible. In addition, the optimal morphological operation size can be taken as 15×15 to obtain a relatively intact plate extraction within a reasonable time.