摘要： 针对传统Chan-Vese(CV)模型对公路路面破损图像分割的局限性，将图像梯度信息引入CV模型，利用路面破损区域纹理与背景的不同，对图像进行分割。引入梯度阈值，将图像的灰度信息和纹理信息相结合，从而使分割方法更具灵活性。实验结果表明，改进的CV模型比传统CV模型具有更好的分割效果。

关键词: 图像分割, 图像梯度, Chan-Vese模型, 路面破损

Abstract: Aiming at the limitations of the traditional Chan-Vese(CV) model applied to the highway pavement disease image segmentation. An improved CV model based on the gradient information is proposed in this paper, which can successfully segment the region with different texture caused by road damage. The threshold of the gradient is introduced to the model which combines the intensity information and the texture information, and increases the flexibility of the method. Experimental results show that the segmentation effect of the improved CV model is better than the traditional CV model.

Key words: image segmentation, image gradient, Chan-Vese(CV) model, pavement diseases

中图分类号: 

• TP391