基于背景色彩和PCNN的磨粒图像单通道分割

本资料包含pdf文件1个，下载需要1积分

为有效地从磨粒图像中分离出磨粒区域，根据彩色磨粒图像背景相对单一的特点，提出了一种基于背景色彩和脉冲耦合神经网络（PcNN）的磨粒图像单通道分割方法。首先，利用Otsu算法对磨粒图像进行预分割处理，并计算图像背景色彩的统计特征，以此作为背景的特征色彩矢量；然后，引入综合色距函数，通过计算图像中各像素与背景特征色彩之间的色差，将彩色磨粒图像三通道问题转化成单通道问题；最后，利用简化PCNN对构造的色差矩阵进行分割，从而得到彩色磨粒图像的分割结果。实验结果表明，该方法对含有单个磨粒和多个磨粒的彩色磨粒图像均能进行精确自动分割，是一种比较理想的分割方法。 The background color of the debris image is quite simple. In order to separate wear particle areas from debris images, a single-channel segmentation method for debris images based on background color and pulse coupled neural network（PCNN） was proposed in this paper. Firstly, Otsu method was utilized for pre-segmentation of debris im- ages. A kind of statistical feature of background color was calculated, which was used as the feature color vector of back- ground. Then, a synthesis color distance function was introduced to compute the color difference between each pixel and the corresponding feature color vector. Thus, the three-channel issue for color debris images was converted into a sin- gle-channel issue. Finally, the constructed color difference matrixes were segmented by simplified PCNN. In this way, the segmentation results of color debris images were achieved. The experiment results prove that the proposed method can accurately and automatically segment the color debris images with either single or many wear particles. It is a perfect segmentation algorithm for color debris images.