Automatic license plate recognition system is an image processing technology used to identify vehicles by their license plates. Such systems require the localization of number plate area in order to identify the characters present on it. Since number plate standards are not strictly practiced in India, a large amount of variations are obtained in parameters like, size of number plate and characters, location of number plate, type of font used, background and foreground colour, etc. which makes the task of number plate localization all the more difficult. The proposed algorithmconsists of two main modules: image splitting and localization of region of interest (ROI). A two-step approach for localization is presented. In the first step, image is split in two parts. In the second step, a two dimensional digital filter is used to find the probable number plate location. This approach enables the localization of number plates in widely varying illumination conditions with relevance to the commonly found types of car number plates with a success rate of 89%.