摘要
Hough变换圆检测(HTCD)存在计算量大和占用内存空间大的缺点,严重影响了该方法的计算效率。提出了一种新的快速Hough变换圆检测(FHTCD)的新方法。首先通过下采样,得到缩小的图像;进行Hough变换圆检测,得到粗尺度的检测结果。将结果恢复到原始尺度,在其放宽的误差范围内建立新的参量空间;再次进行Hough变换,可以得到圆检测精确值。新的方法通过压缩和适度控制参量空间,大大减少了冗余信息带来的计算量,提高了检测效率。实验证明,新的方法计算所需的时间最多不到HTCD方法的1/15;在强抗干扰条件下,其检测速度优于随机圆检测方法,有较好的检测精度,能满足实时处理的要求。
Hough transform circle detection(HTCD)needs amount of calculation and take up a large amount ofmemory.In order to solve these problems,a new method of fast Hough transform circle detection(FHTCD)is proposed.A reduced image can be obtained by down-sampling firstly.HTCD is used to get coarse scale results with the reducedimage;Restoring those results to original scale and establish a new parameter space in their relaxed error range;Accurate value of circle detection can be obtained through HTCD again.The parameter space in the new method is reasonably controlled so as to greatly reduce calculation amount and improved the efficiency of detection.Experiences showthat the time consumption of the new method no more than 1/15 compare with HTCD;Its detection efficiency issuperior to randomized Hough transformation under the condition of strong interferences.The new method has gooddetection precision and can meet the requirements for real time processing.
出处
《电子测量与仪器学报》
CSCD
2010年第9期837-841,共5页
Journal of Electronic Measurement and Instrumentation
基金
国家重点实验室基金(编号:9140C1406020708)
总装基金(编号:9140A01020507DZ02)
粤港重点领域突破项目(编号:20091683)
湖南省科技计划项目(编号:2009FJ3187)
关键词
HOUGH变换
随机HOUGH变换
随机采样
圆检测
参量空间
Hough transformation
randomized Hough transformation
randomized sampling
circle detection
Parameter space