基于灰模型白化响应的边缘检测算法研究

边缘检测是图像分析与处理的重要基础，一直是视觉领域研究的活跃课题。本文深入研究了一种新的基于GM(1,1,C)模型白化响应的图像边缘检测算法。该算法用原图相邻的若干象素点，构建GM(1,1,C)模型，算出相应象素点的白化值，得到原图像素点亮度值与相应白化值之间的误差。依据边缘像素点其亮度与非边缘像素点亮度相差大而不满足GM(1,1,C)建模条件,从而导致边缘像素点白化值误差大的特点，实现边缘检测。实验证明该算法的有效性和具有一定的抗噪能力。

Research on Edge Detection Algorithm Based on Grey Model Whiteriztion Response

Edge detection is the important base of image analysis and processing, which is always the active research subject in vision. A novel edge detection algorithm based on GM(1,1,C) whiterization response was discussed in detail in the paper. The algorithm first built GM(1,1,C) model using neighboring pixels of original image, and calculated the corresponding whiterization response value. So the error between the whiterization value and the original pelses was obtained. As the edge pixel value varies from non-edge pels value largely, which resulted that it can’t satisfy the condition of building GM(1,1,C) and the error of GM(1,1,C) whiterization value is large. According to the error, the edge was deleted. The validation of the algorithm has been proved by test, and the ability of anti-noise also has been tested.