具有抗噪能力的图像阈值分割法

Automatic Thresholding of Gray-Level Image Based on Two-Dimensional Correlation Criterion

图像阈值分割在图像分析和图像识别中具有重要意义 .本文将 Brink所提出利用一维相关系数对图像进行分割的方法推广到二维 .该方法除了考虑系数的灰度信息外 ,还考虑了像素点与其邻域的空间相关信息 .理论与实践表明 ,该文方法在进行图像分割时有很强的抑制噪声性能 。

Image thresholding is an important method for image analysis and recognition. This paper proposes a method of gray scale image thresholding based on two dimensional correlation criterion. The two dimensional correlation criterion utilizes both the gray level information of each pixel and its spatial correlation information within the neighborhood. Comparison between this method and a one dimensional correlation criterion method shows that the proposed method performs much better when images are corrupted by noise.