基于灰度分布的图像三角网格化算法

Image Representation Using Triangular Mesh Based on Content of Image

借助于小波图像分解,提出一种基于图像内容的三角网格表示方法——基于双向模板的图像三角网格化算法.算法考虑图像的灰度分布,利用小波的图像分解能够将图像的各个方向的细节表现出来这一特性,给出符合原始图像灰度分布的三角划分,再对图像的三角划分进行三角网格化,最后获取整幅图像的网格划分.为了得到更好的重建图像质量,对该初始网格进行了细分,并针对三角网格规模的减小做出优化算法.同时提出一种记录模板号和细分点的数据存储结构,用二进制数据流来存储三角网格.通过实验数据对比,该算法能够很好的表示图像,在三角网格规模以及重建图像质量上较其它算法都有一定的优势,是一种极其有效的图像表示方法.

With wavelet decomposition,this paper proposed a method using triangular mesh to represent image base on content of image.With respect to the distributing of gray scale,this algorithm uses seven initial regular triangle templates to represent the image.And then the triangulation partition can be transformed into initial triangulate mesh.The initial triangulate mesh can be subdivided to improve the quality of reconstructed image and raise an optimization algorithm reduce the size of the triangular mesh.A data storage structure is proposed to store the triangular mesh by recording the number of the template and the breakdown points.The experimental results show that： Compared with other image triangulation algorithm,the size of the triangular mesh and the quality of reconstructed image using this algorithm has advantages.