摘要
基于内容的图像检索系统中需要提取图像的形状特征,通常情况下为得到形状特征,需要对图像分割来得到的目标边界。在对不均匀图像提取边界曲线时,噪声点和非噪声区域常会干扰导致小的边界变化。本文采取了一定的措施对该问题进行了改进:首先通过去掉图像的彩色成分,将图像进行灰度化处理,再通过图像增强技术,对图像进行平滑处理,然后提取图像的形状。由于图像增强对图像预制模糊时,已经对图像中的噪声点和颜色不均匀区域进行了处理,所以再次对图像进行形状处理时,可有效去除噪声点和非噪声干扰区域的影响。
Extracting the shape characteristics of images is necessary in the content based image retrieval system. Under normal circumstances, in order to obtain the shape feature, it is need to obtain the target boundary of the image segmentation. When extracting the boundary curve, the area of noise and non-noise constant interference always leads to small boundary changes. In this paper, certain measures have been taken to improve the method of solving the problems. First, by removing the color components of the image, the images are processed with gradation. And then through image enhancement techniques, the smoothing process is done on the image, and the shape of the image is extracted. As the image enhancement technology blurred images by some prefabricated way, the image noises and color uneven region have been processed, which can be used to extract the shape of the images and remove the noise point and non-noise area effectively.
基金
北京高校人才强教资助项目[PHR201106133]
教育教学-本科生科学研究计划项目[PXM2012_014224_000055]
关键词
图像检索
形状特征
噪声
图像复原
维纳滤波
image retrieval
shape characteristic
noises of images
image restoration
Wiener filter
