【24h】

Fast image processing on chain board of inverted tooth chain

机译:倒齿链链板上的快速图像处理

获取原文

摘要

Discussed ordinary image processing technology of inverted tooth chain board, including noise reduction, image segmentation, edge detection and contour extraction etc.. Put forward a new kind of sub-pixel arithmetic for edge orientation of circle. The arithmetic first did edge detection to image by Canny arithmetic, so as to enhance primary orientation precision of edge, then calculated gradient direction, and then interpolated gradient image (image that was detected by Sobel arithmetic) along gradient direction, last obtained sub-pixel orientation of edge. Performed two kinds of least-square fitting methods for line edge to getting its sub-pixel orientation, from analysis and experiments, the orientation error of improved least-square linear fitting method was one quarter of ordinary least-square linear fitting error under small difference of orientation time. The sub-pixel orientation of circle made resolution of CCD increase 42 tines, which enhanced greatly orientation precision of image edge. For the need of quick on-line inspection next step, integrated the whole environment containing image preprocess, Hough conversion of line, setting orientation & direction of image, sub-pixel orientation of line and circle, output of calculation result. The whole quick processing course performed without operator, processing tine of single part is less than 0.3 second. The sub-pixel orientation method this paper posed fits precision orientation of image, and integration calculation method ensure requirement of quick inspection, and lays the foundations for on-line precision visual measurement of image.
机译:讨论了倒齿链板的普通图像处理技术,包括降噪,图像分割,边缘检测和轮廓提取等。提出了一种新型的圆弧边缘方向的亚像素算法。该算法首先通过Canny算法对图像进行边缘检测,以提高边缘的初次定位精度,然后计算梯度方向,然后沿梯度方向插值梯度图像(由Sobel算法检测到的图像),最后获得子像素。边缘的方向。对线边缘进行了两种最小二乘拟合方法以得到其子像素方向,从分析和实验来看,改进的最小二乘线性拟合方法在很小的差异下的定向误差是普通最小二乘线性拟合误差的四分之一。定位时间。圆的亚像素取向使CCD的分辨率提高了42齿,极大地提高了图像边缘的取向精度。为了下一步快速在线检查的需要,集成了包括图像预处理,线的霍夫转换,设置图像的方向和方向,线和圆的子像素方向,计算结果的输出在内的整个环境。整个快速加工过程无需操作员即可完成,单个零件的加工时间不到0.3秒。本文提出的亚像素定位方法适合图像的精确定位,积分计算方法保证了快速检查的要求,为图像的在线精密视觉测量奠定了基础。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号