Image processing algorithm using OpenCV to detect and separate street signs from a picture. The focus of the project was robustness and not speed of processing. The intended use was for creating training data from existing images and not real time applications.