基于人工鱼群霍夫变换的单晶硅直径检测  

Diameter detection of single crystal silicon based on artificial fish swarm algorithm-Hough transform

作　　者：张新雨[1] 刘丁[1] 杨文[1] 杨延西[1] 

机构地区：[1]西安理工大学自动化与信息工程学院,西安710048

出　　处：《仪器仪表学报》2014年第4期940-947,共8页Chinese Journal of Scientific Instrument

基　　金：国家科技重大专项(2009ZX02011001)资助

摘　　要：为提高单晶硅直径检测的速度和精度,首先将单晶硅直径检测转换为在图像中对高亮光环进行椭圆检测的问题,并依据此提出了一种基于人工鱼群霍夫变换的椭圆检测方法,该方法在参数空间中由人工鱼群算法搜索霍夫变换之后使得累加器达到最大值的椭圆参数最优解,从而得到精确的单晶硅直径测量值。最后通过基于带有高斯白噪声的完整椭圆、不完整椭圆的仿真实验和单晶硅直径检测工程实验进行验证,并且和5种传统方法(最小二乘法、霍夫变换、随机霍夫变换、粒子群霍夫变换和RANSAC)进行对比。实验结果表明,所提方法有效地提高了霍夫变换椭圆检测的速度和精度,从而提高了单晶硅直径检测的速度和精度。To improve the speed and accuracy of single crystal silicon diameter detection,the problem of single crystal silicon diameter detection is transformed to the problem of ellipse detection on the highlight halo in the image firstly,and an ellipse detection method based on artificial fish swarm algorithm-Hough transform is proposed.The method uses artificial fish swarm algorithm to search the opti mum ellipse parameters in the parameter space that make the accumulator obtain maximum value after using Hough transform.Then the accurate measurement value of single crystal silicon diameter is obtained.The simulation experiments on the complete and incomplete el lipses with Gaussian white noise and the single crystal silicon diameter detection experiments were carried out.And the results are compared with five kinds of traditional methods （least-square,Hough transform,randomized Hough transform,PSO Hough transform and RANSAC）.Experimental results show that the proposed method effectively improves the accuracy and speed of ellipse detection with Hough transform,thereby improves the accuracy and speed of single crystal silicon diameter detection.

关 键 词：单晶炉 直径检测 椭圆检测 人工鱼群算法 霍夫变换 

分 类 号：TP202.2[自动化与计算机技术—检测技术与自动化装置] TP274.2[自动化与计算机技术—控制科学与工程]