幾種典型邊緣檢測算子的評估

　　圖像信息量巨大，而邊緣信息是圖像的一種緊描述。邊緣是圖像最基本的特征，是用于識別圖像的重要信息。圖像邊緣檢測大幅度地減少了數(shù)據(jù)量，并且剔除了可以認為不相關(guān)的信息，保留了圖像重要的結(jié)構(gòu)屬性。有效的圖像檢測方法對圖像進行處理具有特別重要的意義。該文通過對數(shù)字圖像處理的幾種典型邊緣算子(Robert，Prewitt，Sobel，LOG，Canny)進行評估，并用Matlab工具進行圖像輪廓提取，分析與討論了各種算子的應(yīng)用效果，得出各算子的優(yōu)缺點，為新方法的研究提供衡量尺度和改進依據(jù)。

　　Abstract:

　　Image is enormous amount of information， the edge information is a compact description of the image. Edge is the most basic features of the image and useful information for identifying image. Edge detection greatly reduces the amount of data， and exclude irrelevant information and retain the Important structural properties of image. Effective graphic detection had particular importance to image processing. This paper carried out evaluation through several typical edge operator(Robert，Prewitt，Sobel，LOG，Canny) to digital image and extracted contour with Matlab tools， and analyzed the effect of various operators .We reach advantages and disadvantages of each operator. It provide measurement and improvement for examination of the new method.