基于变形雅可比p=4,q=3-傅里叶矩的纹理图像分割-texture image segmentation based on deformed jacobian p = 4, q = 3 - fourier moments.docxVIP

  • 10
  • 0
  • 约4.7万字
  • 约 49页
  • 2018-05-18 发布于上海
  • 举报

基于变形雅可比p=4,q=3-傅里叶矩的纹理图像分割-texture image segmentation based on deformed jacobian p = 4, q = 3 - fourier moments.docx

基于变形雅可比p=4,q=3-傅里叶矩的纹理图像分割-texture image segmentation based on deformed jacobian p = 4, q = 3 - fourier moments

优秀毕业论文 精品参考文献资料 Texture Segmentation based on Pseudo-Jacobi (p=4,q=3)- Fourier Moments Abstract Texture image segmentati on,which is the base of im age ana lysis and machine visual application ,is an important branch of digital im age processing. It has a very widely application in scientific research and industrial engineering. However,because of two reasons that natural textures have various types,shapes,complex struc阳res and the researchers are lackin g kJ10wledge of the human visual system on the textural mechanism , texture im age segmentation is a crucial problem in the digital image processing field even today. In the past ,researchers have proposed a large number of texture image segmentation algorithms,but these algorithms have some deficiencies . So far,texωre image segmentation is still a challenging task. To probe another belter texture im age segmentation method,a n ew texture image segmentation based on Pseudo -Jacobi (p=4,q=3)-Fouri er m oments is proposed in this paper The research fírstly focuses on th e improved algorithm th at can eliminate coordinate transformation eηor by calculating Pseudo-Jacobi-Fourier moments in Cartesian coordinate svstem directlv. Throu 由deeolv studies on the selection of the feature vector and the width of windows around each pixel,the algorithm of computing moment for each pixel is chosen well and text ure tèat ure of texture image obta ined by u sing Pseudo -Jacobi -Fourier moments. Then classify each pixel íJ1 the textural image using the clustering algorithm . At the same time,compared with the method based on Zemike moments and Orthogonal-Fourier-Me llin moments,method based on Pseudo-Jacobi(p=4,q=3)-Fourier moments has a belter segm entati on performance. Meanwhile ,segmen tation standard for 15 pairs from Brodatz album is fOlmd with segmentation rate of 86.7%. Each texture image used in this paper comes from Brodatz album . The last part of the paper discusses the applica

您可能关注的文档

文档评论(0)

1亿VIP精品文档

相关文档