Noise Removal and Automatic Segmentation of Scanned Color Map

The segmentation result of scanned color map is always unsatisfied due to it′s grievous noise interference.To solve the problem, a new two- step algorithm is presented in this paper.First,WaveShrink and Wiener smoothing is adapted as the main technique to remove noise.Then a concise color clustering algorithm based on probability analysis of the CIE- L~*a~*b~* color space is exploited.By contrast with the traditional segmentation algorithm, this one provides more robustness to noise, satisfying segmentation effect and faster processing velocity.