Image features detection using phase congruency and its application in visual servoing

Image features represent inputs for different types of visual applications. For visual servoing tasks the chosen features must be stable, accurate and robust. Gradient based algorithms for feature detection are sensitive to noise, illumination change and scale change. Using phase congruency, an image feature detection algorithm is developed based on the Local Energy Model. The performances of the detection algorithm were tested using a simulated visual servoing architecture. Simulation was performed using an image sequence acquired by an eye-in-hand real system of a six degree of freedom robot manipulator.