Momentum based level set segmentation for complex phase change thermography sequence

摘要

Level set methods are often used to solve the image segmentation problem. A new method based on momentum level set segmentation is applied to extract phase change line automatically from complex phase change thermography sequence. Firstly, effective phase change information is obtained by the inter-frame difference with the template image and the differential image. Then, in consideration of the relevance of inter-frame phase change line, we present a tracking algorithm of the phase change line by level set segmentation method based on momentum. This new method is derived from gradient descent algorithm. It avoids many unwanted local optima and poor convergence which common level set methods suffer. Moreover, this method greatly improves convergence rate in comparison with C-V model method. The experimental results show that the method is able to solve multi-phase change lines and topology changes problems effectively which can not be solved by existing methods, and can adapt to variant test models.