摘要
由于光照不均、路面污渍及沥青颗粒等因素的影响,传统阈值分割算法应用于路面裂缝处理时效果不佳。针对这一问题提出了基于霍夫变换的OSTU改进算法,对分割后图像进行霍夫变换得到霍夫矩阵,统计其极值和最值分布情况作为判定依据,不断修正由传统OSTU算法得到的初始阈值并重新分割检验,最终得到最佳阈值。对新算法进行理论和实际验证,结果表明,新算法分割效果良好、稳定,抗噪声能力强,极大地提高了裂缝提取的准确度和完整度。
Traditional OSTU algorithm is not applicative in binary process due to the uneven illumination, stains and asphalt particles on surface. Aiming at this problem, an improved algorithm based on hough transformation is proposed, which uses the distribution of the maximum in hough matrix ＇after image segmentation as a judge to correct the initial threshold of traditional OSTU, until find the appropriate threshold. New algorithm overcomes the drawbacks of traditional algorithm. The segmentation results are stable and have strong anti-noise capacity, improve the accuracy and integrity of crack segmentation efficaciously.
出处
《电子设计工程》
2016年第5期43-46,共4页
Electronic Design Engineering
关键词
裂缝检测
图像分割
霍夫变换
霍夫矩阵
crack detection
image segmentation
hough transformation
hough matrix
