An improved CLG Algorithm based on HSV color model

The implementation and optimization of the traditional CLG (Combined local and global) algorithm is always proposed in the gray model. It doesn´t make full use of color information. The same global smoothness factor doesn´t facilitate discontinuity preservation. This paper, firstly, improves the traditional Gradient Constraint Equation; then combines Horn Schunk algorithm and Lucas Kanade algorithm in the HSV color space and chooses adaptive smoothing factor by canny edge detection; finally, comparisons with experiments from the literature demonstrate the favorable performance of the proposed methods with more robustness and 100% density.