摘要
针对微流控聚合酶链式反应(PCR)芯片荧光图像中反应腔识别与荧光信号提取的难题,提出了一种基于算法集成的微流控PCR芯片荧光信号提取方法。首先提取CCD图像中包含多个反应腔的目标区域,采用最小误差阈值分割法分离背景和反应腔,使用网格划分法提取单个反应腔区域,通过小波变换和改进的Canny算子相结合的方法实现反应腔边缘识别。然后对每个反应腔的识别区域进行像素灰度值求和,计算均值来表征本次循环荧光信号的强度。以人工质粒PUC-18基因为检测对象,开展了反应条件相同的四腔微流控实时荧光PCR实验。结果表明,本文方法可快速有效识别PCR反应腔并实时提取每个反应腔的荧光信号,绘制的扩增曲线一致性好,避免了亮、暗斑的影响,提高了基因扩增分析的准确性。
To resolve the problems of polymerase chain reaction（PCR）chambers identification and fluorescence signal extraction in microfluidic real-time PCR system,a data extraction method based on algorithm integration is proposed.First,we extract the target region which contains many PCR chambers from the CCD image.Background and the PCR chambers are separated by the minimum error threshold method.The grid method is used to extract a single reaction chamber area.The edge of each reaction chamber is identified with the wavelet transform and the improved Canny operator.Second,the average gray value of each PCR chamber is calculated and is used as the expression of the gene amplification.A chip with four chambers loading with PUC-18gene is tested by our microfluidic real-time system.The results show that our method can identify PCR chambers effectively and calculate the intensity of the fluorescence signal rapidly.The repeatability of PCR amplification curves is improved by eliminating the influence of the bright and dark spots.The accuracy of gene amplification analysis is also enhanced.
出处
《光电子．激光》
EI
CAS
CSCD
北大核心
2014年第6期1229-1236,共8页
Journal of Optoelectronics·Laser
基金
中科院战略性先导科技专项(XDA08040109)
中科院合肥物质科学研究院院长基金(Y23J321121)资助项目
