边界连接算法探测尿沉渣管形边界

图像的边缘信息是图像分割的重要基础．在图像噪音较小的情况下，经典边界算子可以得到较好的处理结果．但是在处理噪音较大的图像时，经典算子就难以满足要求了．在经典算子的基础上，结合边界跟踪和曲线拟合的理论，得到了能较好处理一般质量图像边界的算法．将该算法应用于模拟图像和实际尿沉渣图像管形边界的探测，比较准确地得到了对象的连续边界。

Edge Connecting Arithmetic for Urine Deposit Tube Detecting

The information about the edge of an image is the important base for image segmentation.Classical edge operator can get good result when there is less noise.But if there is much noise,the result of classical edge operator is unsatisfactory.On the basis of classical edge operator theory and combined with edge tracking and curve fitting theory we developed an algorithm which can deal with the edge of ordinary quality image bet- ter.Finally we apply the improved algorithm on urine deposit image and simulated image for edge detecting,and more accurate continuous edge of the object is obtained.