首页|期刊导航|现代制造工程|融合形态学和Franklin矩的亚像素边缘检测方法

融合形态学和Franklin矩的亚像素边缘检测方法OA

A sub-pixel edge detection method that integrates morphological analysis and Franklin moments

中文摘要英文摘要

对精密测量场景中的检测精度和执行效率进行了研究,提出了一种改进多结构元素形态学梯度和Franklin矩的亚像素边缘检测方法.首先,使用双边滤波对图像进行预处理,构建改进多结构元素形态学梯度滤波算子进行边缘粗定位;然后,通过自适应Otsu阈值分割和Franklin矩数学推导,确定亚像素边缘点模板的关键参数;最后,利用粗定位边缘点结合Franklin矩实现亚像素级精确定位.实验表明,该方法检测精度达到10 μm级,执行效率较对比算法效率提升约16.5%,且抗噪性更好,能更好地满足精密测量边缘检测中高精度和高效率的要求.

The detection accuracy and execution efficiency in the precision measurement scenarios were studied,and an improved sub-pixel edge detection method based on multi-structural element morphological gradient and Franklin moment was proposed.Firstly,bilateral filtering was used for image preprocessing,and an improved multi-structural element morphological gra-dient filter operator was constructed for coarse edge location.Then,the key parameters of the sub-pixel edge point template were determined through adaptive Otsu threshold segmentation and Franklin moment mathematical derivation.Finally,the sub-pixel level precise location was achieved by combining the coarse located edge points with Franklin moment.Experiments show that the detection accuracy of this method reaches the 10 μm level,the execution efficiency was improved by about 16.5%compared with the comparison algorithm,and it had better noise resistance,which can better meet the requirements of high accuracy and high ef-ficiency in edge detection of precision measurement.

闫洪波;余高超;赵威;王鑫辉

内蒙古科技大学机械工程学院,包头 014010内蒙古科技大学机械工程学院,包头 014010内蒙古科技大学机械工程学院,包头 014010内蒙古科技大学机械工程学院,包头 014010

信息技术与安全科学

边缘检测亚像素改进多结构元素形态学梯度自适应Otsu阈值Franklin矩

edge detectionsubpixelimproved multi-structural element morphological gradientadaptive Otsu thresholdingFranklin moments

《现代制造工程》 2026 (8)

132-137,6

内蒙古自然科学基金项目(2024MS05010)

10.16731/j.cnki.1671-3133.2026.08.018

评论