基于灰预测模型的边缘检测新方法

提出了一种基于灰预测理论的边缘检测新方法。该方法基于GM(1,1)模型，根据图像不同区域间灰度级有较大变化的特性，由同行(列)待检测像素之前的若干像素灰度值建立灰预测模型，并利用模型预测此像素灰度值，比较预测灰度值与实际灰度值之间的差值，由差值来判断该像素是否为边缘像素。通过对不同类型测试图像利用本算法进行边缘检测效果比较表明，该算法能取得比较好的效果。

Edge Detection Based on Grey Prediction Model

This paper puts forward a novel algorithm of image edge detection based on grey prediction theory.The grey level of different region in same image changed dramatically,so this character can be used to decide whether a pixel is an edge pixel based on GM(1,1) model.This algorithm establishes GM(1,1) model using four pixel grey values before some pixel,then uses this model to forecast that pixel grey value,finally compares predicting gray value and practicality gray value,if the difference is less than threshold value,then this pixel is not an edge pixel otherwise it is an edge pixel.To testify the effect of this algorithm,the paper uses different types image to experiment the new algorithm,the result show this algorithm has some advantage comparing with traditional algorithm.