摘要
提出了一种新的红外图像分割方法。该方法利用形态学方法来处理红外图像。首先进行形态学滤波,对红外目标图像中的噪声和微小的干扰区域进行滤除,接着提出了一种计算红外图像梯度的多尺度算法提取图像形态学梯度,而后分析了图像分形特征估计方法与形态学梯度的关系,提出了一种新的红外图像分形特征估计算法,在此基础上对图像进行分割。实验结果表明,该算法能较好地解决红外图像的分割问题。
An infrared image segmentation algorithm based on mathematical gradient is presented. First, the noise and small interferential regions are removed with morphological filters. Second, a multiscale algorithm for computing morphological gradient images is proposed and the gradients are derived. Third, the fractal features are extracted and the image is segmented by Otsu's method. Experimental results show the approach is suitable to infrared image segmentation.
出处
《信号处理》
CSCD
2004年第1期10-14,共5页
Journal of Signal Processing
