基于图像拼接的表面粗糙度测量方法

Visual Measurement of Surface Roughness Based on Image Stitching Algorithm

针对光切显微镜图像法测量表面粗糙度时取样长度较小,不足以客观表征工件表面质量的问题,提出一种基于图像拼接的表面粗糙度测量方法。根据表面粗糙度序列图像中相邻帧图像的重叠区域构建匹配模板,采用归一化相关算法计算相似度量进行相邻帧图像匹配与拼接,以加权融合算法平滑拼接后图像的重影和接缝。对拼接后的粗糙度图像提取单侧边缘特征,通过最小二乘法拟合最小二乘轮廓中线,建立轮廓算数平均偏差和轮廓最大高度评定模型。实验结果表明,采用拼接算法后取样长度增加了1983.52μm,测量精度平均提高了1.16%。

In order to solve the problem that the sampling length is too small to objectively characterize the surface quality of workpiece when measuring surface roughness by light-cutting microscope image method,a surface roughness measurement method based on image mosaic is proposed.The matching template is constructed according to the overlapping regions of adjacent frame images in surface roughness sequence images.The similarity measure is calculated by normalized correlation algorithm to match and stitch adjacent frame images,and the weighted fusion algorithm is used to smooth the stitching image's ghosting and seam.The unilateral edge features were extracted from the spliced roughness image,and the least squares contour midline was fitted by least squares method to establish the contour arithmetic mean deviation and contour maximum height evaluation model.The experimental results show that the sampling length is increased by 1983.52μm and the measurement accuracy is increased by 1.16%on average.