摘要
通过痰涂片镜检法检测结核抗酸杆菌,是识别和诊断结核病的一种快速、价廉和特定的检测方法。为提高结核病痰涂片显微图像的质量,提出了一种基于多帧不同曝光图像信息融合,扩展被检测显微图像的动态范围,提高检测视野显微图像质量的实现方法。首先在保持CCD曝光时间不变的条件,通过计算机控制圆盘型光学梯度衰减片旋转定位的方法,摄取同一视场多帧不同曝光量的显微图像序列,从而获取场景的高动态范围显微图像信息。然后利用拉普拉斯金字塔算法将显微图像进行分层,采用图像梯度与信息熵相结合对层次图像进行评价,分配相应的合成权重系数,并用拉普拉斯金字塔逆算法,合成一帧高质量的图像。实验表明,所提出方法能有效地表达亮区和暗区的场景信息,增强图像的细节特征,提高痰涂片显微成像的质量,有利于提高显微图像智能检测的精度。
Sputum smear microscopy to detect acid-fast bacilli （AFB） is a rapid, inexpensive, and highly specific tool for identi- fying persons with active tuberculosis （TB）. A novel method based on multi-frame image information fusion is presented by extending the dynamic range of the microscopic image in order to improve the quality of TB sputum smear microscopic image. Maintained the CCD exposure time, a sequence of microscopic images in the same field with different exposure is obtained by rotating disk-type optical attenuation gradient film controlled by computer in order to obtain high dynamic range scene microscopic image information. After that, the Laplacian pyramid algorithm is used to divide the microscopic image into some layers. The weight of each level is designed according to the information of the image which is evaluated by the image grads and entropy. Finally, the inverse Laplacian pyramid algorithm is applied to synthesize a high-quality image. It is proved that the method proposed in this paper can effectively express the image information in the bright areas and dark areas. It can enhance the image details and improve the quality of microscopy imaging system. Therefore, the method is propitious to the intelligent detection of microscopic image.
出处
《应用激光》
CSCD
北大核心
2010年第3期226-231,共6页
Applied Laser
基金
上海市科学技术委员会基础研究重点基金资助项目(项目编号:08JC14002000)
关键词
显微图像
信息融合
光学梯度衰减片
信息熵
图像细节
microscopy image
information fusion
optical attenuation gradient film
entropy
image detail
