A scale and rotation insensitive algorithm for label location and identification

Image recognition is crucial for today's unmanned reconnaissance vehicles. This paper focuses on an image recognition algorithm designed to be insensitive to scaling and rotation. In addition the algorithm runs in real-time and functions well at low resolutions. The problem of locating and identifying toxic waste labels using an unmanned vehicle is discussed and used as an example for how this algorithm works. An analysis of the strong and weak points of the algorithm is made, along with suggestions for future improvements.