Video Processing Algorithms for Identifying Contexts

Automobiles are integral part of modern transportation. Driving is a complex continuous, multitask processing that involves driver's cognition, perception and motor movements. It is important to understand the various aspects of driving for its incorporation in autonomous vehicles and intelligent systems. A system proficient of enabling this interaction is known as Advanced Driver System (ADAS). The current paper work involves identification of two contexts i.e. car detection and lane detection As opposed to the discrete values that are obtained in the Classical Hough transform, probabilistic Hough transform produces a continuous function. Probablistic Hough Transforms to identify lanes after essential pre-processing steps have been done on the image based on the color features. A Haar-like feature is used to estimate the pixel intensities of the image by running a rectangular window throughout the image and estimate the pixel intensities of the regions by calculating the difference of the sum of each region. This technique is used in identification of cars on the road after training the images from the Renault Nissan Database and both the contexts were tested with the database videos and images.