Automatic exposure compensation for line detection applications

The detection of lines in an image is an important task. In spite of the numerous research papers that have been published on line extraction, there is a lack of real-world applications relating to coping with illumination changes in scenes. This paper provides an automatic exposure compensation scheme that transforms images under unknown illumination conditions into images that can be used as the best preprocessing data for line detection. Our method is tested on gray level frames captured from the camera where the exposure parameter is changed continuously. Among these frames, the image with the best contrast is selected based on the image entropy. We then apply contrast stretching to transform this poorly illuminated image into one that has better visibility. Afterwards, the Canny edge detection algorithm is applied to obtain the input for the Standard Hough Transform, which is the line extraction algorithm. Furthermore, our system detects lines in real-time, so it is suitable for many real-world applications.