论文部分内容阅读
现代成像制导中所使用的红外图像往往存在着噪声大、目标-背景间灰度差较小、边缘较模糊的特点。这些特点会增加边缘提取的难度,因此必须建立更有效的红外图像边缘提取算法以满足需要。针对这些问题,以噪声Gauss分布模型和噪声特征为基础,建立了新型统计学意义下的红外图像边缘检测法。通过对此方法的概率模型进行分析,可以证明在有较大噪声的情况下,只要边缘处的差分值大于一定的值,就能以较大的概率提取出图像边缘。通过在不同情况下与梯度法的抑噪能力进行对比和分析发现,统计边缘提取法的噪声抑制能力要高于梯度法。在与Sobel模板算子法的红外图像边缘检测结果进行仿真和对比后发现,统计法能对红外图像的目标边缘检测取得良好的结果,并且算法具有快速简单的优点。
The infrared images used in modern imaging guidance often have the characteristics of large noise, less gray difference between the target and the background and more fuzzy edges. These features will increase the difficulty of edge extraction, it is necessary to establish a more effective infrared image edge extraction algorithm to meet the needs. Aiming at these problems, based on the noise Gauss distribution model and noise characteristics, a new statistical edge detection method is established. By analyzing the probabilistic model of this method, it can be proved that the edge of the image can be extracted with a higher probability as long as the difference value at the edge is greater than a certain value under the condition of large noise. By contrasting and analyzing the noise suppression ability of the gradient method in different situations, it is found that the statistical edge extraction method has higher noise suppression capability than the gradient method. The results of simulation and comparison with the Sobel template operator’s infrared image edge detection show that the statistical method can achieve good results for the target edge detection of the infrared image, and the algorithm has the advantages of fast and simpleness.