Performance Analysis of Edge Detection Algorithms for Object Detection in Accident Images

摘要

Identifying the severity of the collision in road accidents helps in suitable extricate operation to be initiated immediately. Detecting the objects in accident images demands analyzing the vehicle type, number of vehicles and the severity of collision. Edge detection is used in identifying various objects such as vehicles and victims by outlining the sudden change of pixel intensity from the given accidental images. Prewitt, Roberts, Canny, Sobel, and Laplacian of Gaussian are the predominant edge detection algorithms in literature. Identifying and applying the suitable edge detection algorithm is critical for efficient object detection in accident images. In this paper, the performance of the aforementioned algorithm was evaluated in terms of Peak Signal-to-Noise Ratio (PSNR) and Mean Square Error (MSE). The experimental results show that the Sobel edge detection algorithm is suitable for edge detection in accident images.