基于人工免疫算法的药液颗粒异物检测方法

Artificial Immune Algorithm Based Particle Detection Method for Liquid in Pharmaceutical Container

为实现药液中可见颗粒异物的智能检测,提出了利用人工免疫算法对其中可能存在的小颗粒异物运动轨迹进行辨识的检测方法。通过设计合理的编码方案,构造出小颗粒运动的假设轨迹作为抗原。利用人工免疫算法对抗原优化,通过克隆、变异寻找亲和力大的抗原作为小颗粒异物运动的可能轨迹。最后利用假设检验对这些可能轨迹进行确认来判断颗粒异物是否存在。试验表明,基于该算法的检测方法对药液中的粒径大于50微米的可见颗粒状异物的误检及漏检率小于3%。

To detect visual particles in liquid in pharmaceutical containers, an intelligent method that makes use of artificial immune algorithm to recognize particle tracks is presented. By designing reasonable coding scheme, we construct hypothetic motion trajectories of these particles, which are regarded as antigens. Then artificial immune algorithm is used to optimize these antigens. Through cloning and mutation, antigens that have maximal affinity are found and regarded as the possible tracks of particles. Finally, hypothesis test is used to confirm these possible motion paths and to determine whether the particles exist. Experiment results show that when the proposed method is used to detect visual particles with diameter larger than 50 micron, the rate of miss detection and false detection is less than 3%.