Roadside accidents are one of the leading cause of deaths in India. Though car accidents are preventable, number of causalities due to car accidents have only seen an increase over the years. This paper looks at one of the promising solutions to this problem – autonomous cars. While self-driving cars dominating the Indian roads is still a distant future, it is important to understand and improve the image preprocessing required in a self-driving car before they are omnipresent in our society. This paper interprets the released government data on roadside accidents in India and studies different types of noise with varying intensity to help choose the most appropriate smoothing filter in spatial domain. The paper then focuses on another important aspect of pre-processing, edge detection and compare the prevalent edge-detection techniques with one another. © 2020 IJSTR.