生物芯片微阵列图像倾斜校正算法研究  

Research of incline correction algorithm in microarray image

作　　者：吴灶全[1] 陈熹[1] 娄艳阳 陈杰[1] 刘正春[1] 

机构地区：[1]中南大学地球科学与信息物理学院生物医学工程研究所,湖南长沙410083

出　　处：《中南大学学报（自然科学版）》2014年第4期1091-1097,共7页Journal of Central South University:Science and Technology

基　　金：国家自然科学基金资助项目(61371042,60801019);国家大学生创新课题资助项目(AL11493);中南大学中央高校基本科研业务费专项资金资助项目(2013zzts252)

摘　　要：针对现有生物芯片图像倾斜校正算法校正精度和运算速度的矛盾问题,提出对矩形样点和圆形样点图像分别采用改进的Hough变换和改进的Radon变换方法进行倾斜校正。改进的Hough变换首先对图像进行二值化,并仅统计矩形样点的上下边缘像素的Hough参数和累计值,然后根据所得的Hough参数空间提取占主导的直线组,最终计算出直线组的平均倾斜角度;改进的Radon变换在提取圆形样点边缘后,对边缘图像进行二级Radon变换,从而得到最佳校正角度。实验证明本文的方法能准确、快速地实现芯片图像的倾斜校正,为生物芯片样点识别奠定基础。Considering the contradiction between analysis precision and analysis speed of the now-being image incline correction methods, novel algorithms based on modified Hough transform and Radon transform were proposed to deal with deflected microarray images consisting of rectangular and circular spots, respectively. After binaryzation of a deflected microarray image, the modified Hough transform was applied to a set of pixels lying in the upper and lower edges of every rectangular spot. The dominant straight lines were obtained according to the Hough parameters and the incline angle was then calculated in accordance with the average slope of lines. The modified Radon transform was applied to extract the edges of circular spots and two-stage Radon transform was performed on the obtained edges to generate a proper incline angle. Experimental results show that the methods proposed can realize incline correction of microarray image with high precision and speed, which can lay the foundation for spot identification of microarray image.

关 键 词：生物芯片 倾斜校正 样点识别 HOUGH变换 RADON变换 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]