Gray-scale even image segmentation method based on level set

The invention discloses a gray-scale uneven image segmentation method based on a level set. According to the invention, the global variance information of the C-V model is reserved; local informationis introduced into an energy functional and is superposed with a global gray average value, so that the algorithm has global and localized effects on the edge of an image with uneven gray, and segmentation failure caused by restriction of expansion force and contraction force of a contour curve at a non-edge due to the uneven gray effect is avoided. A new speed stop operator is provided and introduced into a driving force item, the evolution speed of a level set curve is adjusted in the iteration process, evolution is prevented from falling into a local minimum value, and therefore a smootherevolution curve is obtained, and a smoother and more accurate segmentation result can be obtained by introducing the speed stop operator into the driving force item. 本发明公开了一种基于水平集的灰度不均图像分割方法，本发明在保留C-V模型的全局方差信息的同时，在能量泛函中引入局部信息，