基于改进PCNN和局部相关度的药液检测图像分割

Image segmentation based on modified-PCNN and partial correlation for medical liquid detection

药液检测是医药检测邻域中一个重要的研究课题。利用机器视觉对瓶装药液进行可见异物的检测是药液检测发展的必然趋势。在机器视觉检测中,如何从药液图像中准确提取杂质信号是一个极为重要的环节。该文在分析了脉冲耦合神经网络(PCNN)特点和光流噪声的基础上,提出了一种新的图像分割方法。这种方法将局部相关系数耦合到PCNN的权值连接矩阵中,实现了对药液图像中杂质信号的精确提取,消除了药液检测过程中光流噪声对弱小杂质目标分割的影响。

Medical liquid detection is important in medical facilities. Machine vision is being widely used to find foreign bodies in liquids for medical liquid detection. However, the key step is extracting the impurity information from the liquid medicine images. An image segmentation method based on PCNN and light flow noise is described in this paper. The method uses partial correlation coefficients to construct the weight connection matrix to more accurately extract impurity information and effectively distinguish weak impurities from light flow noise.