摘要
针对图像中Marangoni凝结液边缘由模糊性引起的不确定性问题,以及为后续计算凝结液面积和统计提供科学依据,提出了一种基于改进的Otsu法和四叉树结构相结合的新型边缘检测方法。先自动确定高低阈值分割图像、去噪,采用四叉树法定位凝结液的位置;再考虑到复杂纹理特征,通过逻辑运算进一步处理了噪声,完善了凝结液的完整性。最后用Canny算子获得了很好的凝结液边缘。实验结果表明,对凝结液提取具有较高的区分度且精确性高、鲁棒性较强,降低了计算凝结液半径、覆盖率、换热系数等的复杂性,提高了计算速度,在工业设计方面有着重要的意义。
Due to solve the uncertainty caused by fuzziness of Marangoni condensation drop edge and provide scientific proof for subsequent calculation of drop area and statistics,optimized Otsu combined with quadtree structure was put forward as a new edge detection method. This algorithm segment image by automatically confirming the high-low double thresholds and denoising,positioned the location of condensation by applying the quadtree algorism,considering the complex texture features,dealing with the noise by logical computing,the integrity of condensation edge was improved. Finally a better condensation liquid edge has been obtained through Canny algorithm.The experimental results show that the algorithms get batter differentiating degree,high accuracy and robustness.The complexity of calculation condensation liquid radius and coverage rate,accelerate calculation speed was Reduced in calculating the heat transfer coefficient,essential in the industrial designing.
出处
《科学技术与工程》
北大核心
2017年第8期187-191,共5页
Science Technology and Engineering
基金
新疆维吾尔自治区自然科学基金(2016D01C048)资助
