Color recognize algorithm based on neural network in augment reality

Color recognition is one of the most important issues in the Augmented Reality System. It is based on color-mark registration. The traditional color filter algorithm is easily affected by environment. Because of the complexity of the color structure. The paper illustrates a new method by which more accurate color recognition is achieved. Firstly, the base structure RGB (Red Green Blue) color is changed into HSV (Hue-Saturation-Value) structure. Then utilized neural network to check color by their hues. After that checked whether saturation and lightness are up to the requirement. Using the neural network algorithm, color-marks can be recognized accurately in different illuminations.