面元法的红外制导图像边缘检测

红外图像通常存在着噪声大、目标和背景之间具有较小的灰度差、边缘较模糊的特点,使现有的方法不能有效地提取红外图像的边缘。而图像的表面积特征能较好地针对红外图像的这些特征提取出边缘,因此利用图像的表面积特征提出了一种实用的面元边缘检测法。在对图像灰度值变化情况进行分析的基础上,使用了一种十字面积计算公式以满足边缘检测的需要。通过噪声误差对面元法和梯度法计算结果的影响分别进行分析,发现面元法比梯度法的噪声抑制能力至少高两倍,而且在边缘处的噪声抑制能力更高。将面元法几种常用模板算子法的红外图像边缘检测结果进行实际对比后发现,所提出的面元法对红外图像的目标边缘检测能取得良好的结果。

Edge Detection Investigation of Infrared Image based on Cell Surface Method

Characteristics of IR images mostly include big noise,small distinction betweens target and background,blurry edge and so on.These characteristics make existent methods ineffective and require more effective method.Surface of image can solve the questions better,so a new cell surface method is put forward to detect edge in this paper.By analysis the change of gray value,a cross formula is used to compute surface.After analysis the infection of noise for cell surface method and gradient method,noise restrain capacity of cell surface method is two times higher than that of gradient method.At the same time,we also find that capacity of noise restrain in edge is better.A comparison is made between the cell surface method and four of the most common algorithms.The analyses indicate that good results can be get when this cell surface method is used todetect edge of IR image.