Abstract: A new method for CBIR based on shape Fourier descriptor is presented．The first step of this method is pre-processing the image，then carrying the image segmentation based on topology relationship，and then use contour tracking method to obtain the contour feature which later described by Fourier descriptor using FFT，at last，using block distance to measure the similarity between the original picture and target picture.In order to verify the effectiveness of the method，we experimented under MATLAB and VC++ environment.By several experiments，the results show that this method is effective and also by algorithm analysis，we conclude that the method is rational for image retrieval.

Key words: shape describing, Fourier descriptor, spatial topology, comparability

摘要： 提出了一种利用轮廓特征和拓扑关系的图像检索方法。首先对图像进行预处理，然后根据空间拓扑关系进行区域分割，再对各个区域进行轮廓跟踪，并用傅立叶描述子对所提取区域的轮廓进行描述，最后用街区距离进行相似度的判定。为了验证该算法的有效性，在MATLAB和VC++环境下进行了实验，通过实验结果和算法分析，证明了该方法的合理性和有效性。

关键词: 形状描述, 傅立叶描述子, 空间拓扑关系, 相似度