Content-based image retrieval (CBIR) overcomes the problem that the text-based image retrieval is unable to use rich visual features in image. CBIR more fits the habit of human's cognition. The CBIR has become a hot spot in recent years. Among them, the texture feature is an important visual image features. The experiments, combining the characteristics of cultural relics, turns the colorful cultual relics into grayscale image first. Then detect the edge by Canny factor. And use im proved Sobel operator simulate co-occurrence matrix which tests 4 directions texture matrix. Ultimately convert the image to a histogram to extract the texture features of image artifacts. Experiments showed that texture features are artifacts can get a good image retrieval to retrieve results.%基于内容的图像检索克服了基于文本的图像检索方法无法利用图像中所包含的丰富的视觉特征的问题,更符合人类认识事物的习惯,成为了近些年来图像检索的热点.其中,纹理特征是图像中一个重要的视觉特性.实验结合文物图像特点,首先将彩色文物图像转化为灰度图像,然后用Canny算子检测出图像边缘,再用改进的Sobel算子模拟共生矩阵检测的4个方向的纹理矩阵,最终将其转换为直方图,从而提取到文物图像的纹理特征.实验表明,使用这种纹理特征进行文物图像检索可以取得很好的检索效果.
展开▼
