基于视觉模型的自适应图象边缘检测

An Adaptive Algorithm for Edge Detection Based on Visual Perception Model

本文提出了一个用于图象边缘检测的视觉模型和对数域的最大块梯度（ＬＭＧＧ）算法，它们能自适应地表达人眼在不同背景亮度下的最小分辨力，并以此阀值进行图象的边缘检测。根据该模型和ＬＭＧＧ算法检测的图象边缘较之其它方法与人眼观测到的结果更为一致。实验表明，该方法取得了良好的效果。

This paper presents a visual perception model and an adaptive algorithm called maximum group gradient on the field of logarithm(LMGG) for image edge detection. The given model , as an adaptive threshold, can approximately describe the least visual resolution on the background with various intensities, while the inspection result by the LMGG algorithm has a better consistency with that by human than that by other methods. The experiment indicates that this method is efficient and useful.